COMPARISON OF RECENT MAJOR SPRING FLOOD EVENTS IN
THE SAINT JOHN RIVER BASIN

by

Lisa Rickard
B.Sc., University of New Brunswick, 2016

THESIS SUBMITTED IN PARTIAL FULFILMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
April 2023
© Lisa Rickard, 2023

I

Committee Members
Supervisor: Stephen Déry. PhD
Department of Geography, Earth and Environmental Sciences
University of Northern British Columbia
Prince George, Canada
Supervisor: Ronald Stewart. PhD
Department of Environment and Geography
University of Manitoba
Winnipeg, Canada
Committee Member: Julie Thériault. PhD
Département des sciences de la Terre et de l'atmosphère
Université du Québec à Montréal
Montréal, Canada
Committee Member: Peter Jackson. PhD
Department of Geography, Earth and Environmental Sciences
University of Northern British Columbia
Prince George, Canada

II

Abstract
The flood-prone Saint John River (SJR) traverses provincial and international borders as it
travels from its source in northern Maine to its mouth in southern New Brunswick (NB). In 2008,
NB experienced its worst spring flood in 35 years, which was followed by more major spring
flooding in 2018 and 2019. As part of the Saint John River Experiment on Cold Season Storms
(SAJESS), the objectives of this project are to identify the sequence of events that led to these
floods, and to compare these to the 2021 season, in which no major spring flooding occurred.
Relying largely upon evaluated reanalysis and hydrometric data, numerous atmospheric, surface,
and hydrological variables are examined at various spatial and temporal scales.
There are commonalities and differences between flood years as well as between flood
years and the non-flood year. When averaged across the upper basin, flood years show
consistency in terms of positive winter and spring precipitation anomalies, positive snow water
equivalent (SWE) anomalies, and steep increases in April cumulative runoff. However, they show
inconsistency in terms of ice jams and positive spring total precipitation anomalies when
averaged over the full basin. A comparison of the conditions between flood and non-flood years
also reveals commonalities, such as northeastward-moving storms affecting the region and
positive winter total precipitation anomalies when averaged over the full basin. There are also
differences, such as the early snowmelt and early timing of peak flow and water level in the nonflood year. As well, rain-on-snow events were a prominent feature of the three flood years but
not the non-flood year—not because there were no rainstorms, but because there was low SWE
when they occurred, due to the early snowmelt. Overall, the concurrence, or lack thereof, of key
meteorological-related conditions is a critical issue affecting the likelihood of flooding.
III

Table of Contents
Abstract .....................................................................................................................................III
List of Tables ............................................................................................................................ VII
List of Figures ............................................................................................................................ IX
List of Abbreviations............................................................................................................... XVII
Acknowledgements ............................................................................................................... XVIII
Chapter 1 Introduction ................................................................................................................1
1.1 Territorial Acknowledgement ............................................................................................1
1.2 Motivation.........................................................................................................................1
1.3 Objectives..........................................................................................................................2
1.4 Thesis Structure .................................................................................................................2
Chapter 2 Literature Review ........................................................................................................4
2.1 Physiography and Climatology ...........................................................................................4
2.2 Changing Climate.............................................................................................................10
2.3 Atmospheric Circulations and Storms ..............................................................................11
2.4 Spring River Flooding .......................................................................................................13
2.5 Causative Classification of Floods ....................................................................................15
2.5.1 Hydroclimatic Perspective ........................................................................................16
2.5.2 Hydrologic Perspective .............................................................................................17
2.5.3 Hydrograph Perspective............................................................................................18
Chapter 3 Data and Methods ....................................................................................................19
3.1 Reanalysis Data ...............................................................................................................19
3.1.1 ERA5 and ERA5-Land.................................................................................................19
3.1.2 NCEP-NCAR Reanalysis..............................................................................................20
3.2 Weather Station Data ......................................................................................................23
3.3 Hydrometric Data ............................................................................................................25
3.4 Data Comparison .............................................................................................................30
3.5 Analysis ...........................................................................................................................36
3.5.1 Hydroclimatic Perspective ........................................................................................36
3.5.2 Hydrologic Perspective .............................................................................................39
3.5.3 Hydrograph Perspective............................................................................................41
Chapter 4 Results ......................................................................................................................43

IV

4.1 Hydroclimatic Perspective ...............................................................................................44
4.1.1 Monthly Standardised Geopotential Height Anomaly ...............................................44
4.1.2 Monthly Standardised Sea Level Pressure Anomaly ..................................................52
4.1.3 Daily Mean Geopotential Height at 500 hPa .............................................................55
4.1.4 Storm Tracks .............................................................................................................60
4.2 Hydrologic Perspective ....................................................................................................64
4.2.1 Climate Normals .......................................................................................................64
4.2.2 Monthly Anomalies (Temperature, Precipitation, SWE, and Soil Water) ...................66
4.2.3 Seasonal Precipitation and Temperature Anomalies .................................................76
4.2.4 Rainfall versus SWE...................................................................................................80
4.2.5 Heavy Precipitation Events .......................................................................................83
4.2.6 Antecedent Soil Moisture .........................................................................................88
4.3 Hydrograph Perspective ..................................................................................................94
4.3.1 Flood Hydrograph Patterns .......................................................................................94
4.3.2 Station-Specific Temporal Comparisons ....................................................................99
4.3.3 Lagged Correlations of Hydrometric Gauges ...........................................................101
Chapter 5 Discussion and Synthesis .........................................................................................104
5.1 Flood Years (2008, 2018, 2019)......................................................................................104
5.1.1 Chronology .............................................................................................................104
5.1.2 Seasonal Patterns ...................................................................................................111
5.1.3 Rain-on-Snow .........................................................................................................112
5.2 Non-Flood Year (2021) ...................................................................................................115
5.2.1 Chronology .............................................................................................................115
5.2.2 Seasonal Patterns ...................................................................................................117
5.2.3 Rain-on-Snow .........................................................................................................117
5.3 Comparisons between Flood and Non-Flood Years ........................................................118
5.3.1 Commonalities........................................................................................................118
5.3.2 Differences .............................................................................................................119
5.4 Comparison with Previous Studies .................................................................................121
Chapter 6 Concluding Remarks................................................................................................124
6.1 Summary .......................................................................................................................124
6.2 Future work ...................................................................................................................126
6.2.1 Additional Analyses.................................................................................................126
V

6.2.2 Flood Preparedness and Mitigation ........................................................................127
References ..............................................................................................................................129
Appendices .............................................................................................................................137
A.1 Monthly Geopotential Height Anomaly .........................................................................137
A.2 Monthly Sea Level Pressure Anomaly ............................................................................144

VI

List of Tables
Table 2.1 Description of Saint John River reaches and sub-reaches (data source: Kidd et al., 2011,
table 2.1, p. 24). ..........................................................................................................................7
Table 3.1 Time periods of reanalysis data, including the variables used and the corresponding
results sections in which they appear. As applicable, anomalies of variables were calculated
relative to the 1991 - 2020 period for all reanalysis datasets. ....................................................21
Table 3.2 Weather station metadata (data sources: CoCoRaHS, n.d.; GC, 2021a; NWS, n.d.). ....25
Table 3.3 Hydrometric station metadata, including the parameters used in analyses (26 stations)
(data sources: USGS, 2021; WSC, 2021). ....................................................................................27
Table 3.4 Kling-Gupta efficiency (KGE) values and their corresponding model performance
classifications (Mai et al., 2022).................................................................................................32
Table 3.5 Comparison of mean monthly air temperature (°C) between ERA5-Land and weather
station data, including average bias, average RMSE, and Kling-Gupta efficiency (KGE), over a
reference period of 2011 - 2021. ...............................................................................................33
Table 3.6 Comparison of mean monthly total precipitation (mm) between ERA5-Land and
weather station data, including average bias, average RMSE, and Kling-Gupta efficiency (KGE),
over a reference period of 2011 - 2021. ....................................................................................34
Table 3.7 Comparison of mean monthly (October to May) snow depth (cm) between ERA5-Land
and weather station data, including average bias, average RMSE, and Kling-Gupta efficiency
(KGE), over a reference period of 2011 - 2021. ..........................................................................35
Table 3.8 Comparison of mean monthly (October to May) snow depth (cm) between ERA5-Land
and weather station data, including average bias, average RMSE, and Kling-Gupta efficiency
(KGE), over a reference period of 2015 - 2021. ..........................................................................35
Table 3.9 Comparison of mean monthly (January to April) SWE (mm) between ERA5-Land and
Canadian historical snow survey data, including average bias, average RMSE, and Kling-Gupta
efficiency (KGE), over a reference period of 1974 - 2016. ..........................................................36
Table 4.1 Number of storm tracks that passed south of the Saint John River basin (SJRB), directly
over the SJRB, or north of the SJRB. The two entries separated by a slash are the number from
December (Dec) to May and the number from March (Mar) to May of the four years of interest.
Storm tracks are mapped in Figures 4.13 and 4.14. ...................................................................61
Table 4.2 Heavy precipitation event thresholds (mm), defined by the 90 th percentile values,
specific to the phase (total, liquid, and solid precipitation), season (autumn: September, October,
November; winter: December, January, February; spring: March, April, May), and region (upper,
VII

middle, and lower Saint John River basins). Thresholds were calculated from ERA5-Land data
relative to the 1991 - 2020 period. ............................................................................................83
Table 5.1 Summary of key dates and values for the flood (2008, 2018, 2019) and non-flood (2021)
years. ......................................................................................................................................106
Table 5.2 Key features of the heaviest rainfall events from 1 April - 15 May for the flood (2008,
2018, 2019) and non-flood (2021) years. .................................................................................114
Table 5.3 Contributing factors during the flood (2008, 2018, 2019) and non-flood (2021) years.
...............................................................................................................................................120

VIII

List of Figures
Figure 2.1 Reaches and sub-reaches of the Saint John River, including start and end points (black
dots). The shaded polygons identify the upper (yellow), middle (red), and lower (green) basins,
and the numbered white circles refer to the sub-reaches described in Table 2.1. Political
boundaries are shown with a green line (image source: Kidd et al., 2011, figure 2.1, p. 25). .......6
Figure 2.2 Elevation of the Saint John River basin, including rivers and main stem hydroelectric
dams (data source: United States Geological Survey [USGS], 2013). ............................................8
Figure 2.3 Ecological regions of the Saint John River basin. The seven ecoregions in the Canadian
portion (red shades) include: Appalachians (117), Northern New Brunswick Highlands (118), New
Brunswick Highlands (119), Saint John River Valley (120), Southern New Brunswick Uplands (121),
Maritime Lowlands (122), and Fundy Coast (123). The five biophysical regions in the American
portion (purple shades) include: Boundary Plateau (1), Saint John Uplands (2), Aroostook Hills (3),
Aroostook Lowlands (4), and Eastern Interior (183) (data sources: GC, 2021b; Maine GeoLibrary,
2017)...........................................................................................................................................9
Figure 2.4 Typical winter mid-latitude cyclone paths, with the location of the Saint John River
basin identified (red box) (image source: Ahrens et al., 2016, figure 12.5). ...............................12
Figure 2.5 Daily water level above the local datum at the Fredericton hydrometric station for the
2008, 2018, and 2019 water years, with the dashed red line indicating the flood stage (data
source: Water Survey of Canada [WSC], 2021). .........................................................................15
Figure 3.1 Location of weather stations and snow survey sites that were used to assess the
accuracy of the ERA5-Land product within the Saint John River basin, with symbols identifying
the respective collecting agencies, including the Community Collaborative Rain, Hail and Snow
Network (CoCoRaHS) (CoCoRaHS, n.d.), Environment and Climate Change Canada (ECCC) (GC,
2021a), the National Weather Service (NWS) (NWS, n.d.), and the Canadian Historical Snow
Survey (CHSS) (GC, 2021c). Weather station details are provided in Table 3.2. ..........................24
Figure 3.2 Location of hydrometric data sources, indicating water level (m) or flow (m3 s-1) (data
sources: USGS, 2021; WSC, 2021). Locations of main stem hydroelectric generating stations are
also identified. Hydrometric station details are provided in Table 3.2. ......................................26
Figure 3.3 Comparison of mean monthly air temperature (°C) between ERA5-Land (black lines)
and weather stations operated by ECCC (red lines) and NWS (blue lines), over a reference period
of 2011 - 2021. ..........................................................................................................................32
Figure 3.4 Comparison of mean monthly total precipitation (mm) between ERA5-Land (black
lines) and weather stations operated by ECCC (red lines) and NWS (blue lines), over a reference
period of 2011 - 2021. ...............................................................................................................33

IX

Figure 3.5 Comparison of mean monthly snow depth (cm) from October to May between ERA5Land (black lines) and weather stations operated by ECCC (red lines) and NWS (blue lines), over
a reference period of 2011 - 2021. ............................................................................................34
Figure 3.6 Comparison of mean monthly (October to May) snow depth (cm) between ERA5-Land
(black lines) and weather stations operated by CoCoRaHS (green lines), over a reference period
of 2015 - 2021. ..........................................................................................................................35
Figure 3.7 Comparison of mean monthly (January to April) SWE (mm) between ERA5-Land (black
lines) and Canadian historical snow survey data (green lines), over a reference period of 1974 2016. .........................................................................................................................................36
Figure 4.1 December standardised geopotential height anomalies (at 250, 500, and 925 hPa) for
the four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 60 geopotential metres (gpm).
Positive standardised geopotential height anomalies are shown in yellow and orange, and
negative anomalies are shown in shades of blue, with shading starting at ±1 standard deviation
and a contour interval of 0.5 standard deviation. Anomalies are calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). .......................................46
Figure 4.2 January standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the
four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 60 geopotential metres (gpm).
Positive standardised geopotential height anomalies are shown in yellow and orange, and
negative anomalies are shown in shades of blue, with shading starting at ±1 standard deviation
and a contour interval of 0.5 standard deviation. Anomalies are calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). .......................................47
Figure 4.3 February standardised geopotential height anomalies (at 250, 500, and 925 hPa) for
the four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 60 geopotential metres (gpm).
Positive standardised geopotential height anomalies are shown in yellow and orange, and
negative anomalies are shown in shades of blue, with shading starting at ±1 standard deviation
and a contour interval of 0.5 standard deviation. Anomalies are calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). .......................................48
Figure 4.4 March standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the
four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 60 geopotential metres (gpm).
Positive standardised geopotential height anomalies are shown in yellow and orange, and
negative anomalies are shown in shades of blue, with shading starting at ±1 standard deviation
and a contour interval of 0.5 standard deviation. Anomalies are calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). .......................................49

X

Figure 4.5 April standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the
four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 60 geopotential metres (gpm).
Positive standardised geopotential height anomalies are shown in yellow and orange, and
negative anomalies are shown in shades of blue, with shading starting at ±1 standard deviation
and a contour interval of 0.5 standard deviation. Anomalies are calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). .......................................50
Figure 4.6 May standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the
four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 60 geopotential metres (gpm).
Positive standardised geopotential height anomalies are shown in yellow and orange, and
negative anomalies are shown in shades of blue, with shading starting at ±1 standard deviation
and a contour interval of 0.5 standard deviation. Anomalies are calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). .......................................51
Figure 4.7 Monthly (December, January, and February) average sea level pressure standardised
anomalies for the four water years of interest, with respect to the 1991 - 2020 reference period.
Standardised anomalies are calculated from the NCEP-NCAR Reanalysis dataset and mapped in
an interval of 0.5 standard deviation (image source: IRI, n.d.). ..................................................53
Figure 4.8 Monthly (March, April, and May) average sea level pressure standardised anomalies
for the four water years of interest, with respect to the 1991 - 2020 reference period.
Standardised anomalies are calculated from the NCEP-NCAR Reanalysis dataset and mapped in
an interval of 0.5 standard deviation (image source: IRI, n.d.). ..................................................54
Figure 4.9 Daily mean geopotential height in metres at 500 hPa during the spring 2008 peak water
level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 30 April - 4 May (five
days). Maps were produced with R (version 4.1.0) using the ERA5 dataset. ..............................56
Figure 4.10 Daily mean geopotential height in metres at 500 hPa during the spring 2018 peak
water level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 30 April 30 - 7
May (eight days). Maps were produced with R (version 4.1.0) using the ERA5 dataset. .............57
Figure 4.11 Daily mean geopotential height in metres at 500 hPa during the spring 2019 peak
water level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 23 April - 26
April (four days). Maps were produced with R (version 4.1.0) using the ERA5 dataset. .............58
Figure 4.12 Daily mean geopotential height in metres at 500 hPa during the spring 2021 peak
water level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 3 - 11 April
(nine days). Maps were produced with R (version 4.1.0) using the ERA5 dataset. .....................59
Figure 4.13 Monthly low pressure system tracks for December (Dec), January (Jan), and February
(Feb) generated from ERA5 data for the four water years of interest, with first points identified

XI

by the code represented by black dots and the Saint John River basin delineated by the black
polygon. ....................................................................................................................................62
Figure 4.14 Monthly low pressure system tracks for March (Mar), April (Apr), and May generated
from ERA5 data for the four water years of interest, with first points identified by the code
represented by black dots and the Saint John River basin delineated by the black polygon. ......63
Figure 4.15 Monthly climate normals of air temperature, precipitation, and its phase for the Saint
John River basin (SJRB) using the ERA5-Land dataset, in reference to the 1991 - 2020 period.
Normals are shown for the full, upper, middle, and lower SJRB.................................................65
Figure 4.16 Mean monthly air temperature (T) anomalies for the full, upper, middle, and lower
Saint John River basins for the four water years of interest. Anomalies were derived from the
ERA5-Land dataset in reference to the 1991 - 2020 period........................................................67
Figure 4.17 Mean monthly total precipitation (TP) anomalies for the full, upper, middle, and lower
Saint John River basins for the four water years of interest. Anomalies were derived from the
ERA5-Land dataset in reference to the 1991 - 2020 period........................................................69
Figure 4.18 Mean monthly liquid precipitation (LP) anomalies for the full, upper, middle, and
lower Saint John River basins for the four water years of interest. Anomalies were derived from
the ERA5-Land dataset in reference to the 1991 - 2020 period. ................................................70
Figure 4.19 Mean monthly solid precipitation (SP) anomalies for the full, upper, middle, and lower
Saint John River basins for the four water years of interest. Anomalies were derived from the
ERA5-Land dataset in reference to the 1991 - 2020 period........................................................71
Figure 4.20 Mean monthly snow water equivalent (SWE) anomalies for the full, upper, middle,
and lower Saint John River basins for the four water years of interest. Anomalies were derived
from the ERA5-Land dataset in reference to the 1991 - 2020 period. ........................................73
Figure 4.21 Mean soil water (SW) anomalies for the full, upper, middle, and lower Saint John River
basins for the four water years of interest. Anomalies were derived from the ERA5-Land dataset
in reference to the 1991 - 2020 period. .....................................................................................75
Figure 4.22 Standardised winter (December, January, and February) precipitation and air
temperature anomalies for the full, upper, middle, and lower Saint John River basins using the
ERA5-Land dataset, in reference to the 1991 - 2020 reference period. The four years of interest
are identified with white dots. Years that fall outside of the hashed box have a difference greater
than two standard deviations from the reference period. .........................................................77
Figure 4.23 Standardised spring (March, April, and May) precipitation and air temperature
anomalies for the full, upper, middle, and lower Saint John River basins using the ERA5-Land
dataset, in reference to the 1991 - 2020 reference period. The four years of interest are identified
with white dots. Years that fall outside of the hashed box have a difference greater than two
standard deviations from the reference period. ........................................................................79
XII

Figure 4.24 Daily total rainfall and mean snow water equivalent (SWE) from 1 April - 15 May for
the four years, calculated for the full, upper, middle, and lower Saint John River basins using the
ERA5-Land dataset. ...................................................................................................................81
Figure 4.25 Cumulative (cum) rainfall and snowmelt from 1 April - 15 May for the four years,
calculated for the full, upper, middle, and lower Saint John River basins using the ERA5-Land
dataset. .....................................................................................................................................82
Figure 4.26 The total number of heavy daily precipitation (total, liquid, and solid) events per
season for the four years of interest, specific to the upper (top row), middle (middle row), and
lower (bottom row) Saint John River basin. ...............................................................................85
Figure 4.27 Heavy daily precipitation (total, liquid, and solid) anomalies per season for the four
years of interest, specific to the upper (top row), middle (middle row), and lower (bottom row)
Saint John River basin................................................................................................................86
Figure 4.28 Heavy daily precipitation (total, liquid, and solid) positive (pos) and negative (neg)
anomalies per season for the four years of interest, over the full Saint John River basin. ..........87
Figure 4.29 Daily total runoff and mean soil water content from 1 April - 15 May for the four years
of interest calculated for the full, upper, middle, and lower Saint John River basins using the ERA5Land dataset. ............................................................................................................................90
Figure 4.30 Cumulative (cum) runoff from 1 April - 15 May for the four years of interest calculated
for the full, upper, middle, and lower Saint John River basins using the ERA5-Land dataset. .....91
Figure 4.31 Spatial distribution of soil water (SW) (liquid and solid) anomalies for winter
(December, January, and February) and spring (March, April, and May) for the four years of
interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020
period. ......................................................................................................................................92
Figure 4.32 Spatial distribution of soil water (SW) (liquid and solid) (top row) and 2 m air
temperature (2m T) (bottom row) anomalies in April for the four years of interest. Anomalies
were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period....................93
Figure 4.33 Flood hydrographs for 1 April - 15 May 2008 derived from 26 hydrometric stations
across the Saint John River Basin. The upper, middle, and lower sections of the river are
delineated by the thick black lines.............................................................................................95
Figure 4.34 Flood hydrographs for 1 April - 15 May 2018 derived from 26 hydrometric stations
across the Saint John River Basin. The upper, middle, and lower sections of the river are
delineated by the thick black lines.............................................................................................96
Figure 4.35 Flood hydrographs for 1 April - 15 May 2019 derived from 26 hydrometric stations
across the Saint John River Basin. The upper, middle, and lower sections of the river are
delineated by the thick black lines.............................................................................................97
XIII

Figure 4.36 Hydrographs for 1 April - 15 May 2021 derived from 26 hydrometric stations across
the Saint John River Basin. The upper, middle, and lower sections of the river are delineated by
the thick black lines. ..................................................................................................................98
Figure 4.37 Hydrometric station-specific temporal comparisons at five locations along the main
stem of the Saint John River, for the four years of interest. For stations reporting water level, the
flood stage is indicated by a horizontal dashed line. The upper, middle, and lower sections of the
river are delineated by the thick black lines.............................................................................100
Figure 4.38 Lag time correlations (0 - 10 days) between eight main stem hydrometric gauges and
water level at the Saint John, NB gauge (near the Saint John River estuary), including the longterm average (1991 – 2020 reference period) and from 1 April to 15 May for the four years of
interest. Gauges are ordered from farthest (Ninemile) to closest (Oak Point) to the Saint John
gauge. Maximum lag values are labelled by the corresponding integer, and red circles identify
correlation coefficients that are not statistically significant (p-value ≤ 0.05)............................103
Figure 5.1: Travel times of peak flow or water level for the 2008, 2018, and 2019 spring floods
along the Saint John River main stem, from Ninemile Bridge (plotted as zero) in the headwaters
to Saint John at the estuary. Main stem hydroelectric dams are labelled and identified by vertical
lines. .......................................................................................................................................107
Figure 5.2 Heavy rainfall events and related storms for 1 April - 15 May 2008. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.....................................................................................................................108
Figure 5.3 Heavy rainfall events and related storms for 1 April - 15 May 2018. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.....................................................................................................................109
Figure 5.4 Heavy rainfall events and related storms for 1 April - 15 May 2019. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
XIV

line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.....................................................................................................................110
Figure 5.5 Heavy rainfall events and related storms for 1 April - 15 May 2021. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.....................................................................................................................116
Figure A.1 December geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) of the
four water years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology
values are shown as grey contours with a contour interval of 30 geopotential metres (gpm).
Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies
are shown in shades of blue and are also expressed in units of gpm. Anomalies are calculated
from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). ...138
Figure A.2 January geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four
years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology values are
shown as grey contours with a contour interval of 30 geopotential metres (gpm). Positive
geopotential height anomalies are shown in yellow and orange, and negative anomalies are
shown in shades of blue and are also expressed in units of gpm. Anomalies are calculated from
the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI Climate and
Society Map Room, n.d.). ........................................................................................................139
Figure A.3 February geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the
four years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology values are
shown as grey contours with a contour interval of 30 geopotential metres (gpm). Positive
geopotential height anomalies are shown in yellow and orange, and negative anomalies are
shown in shades of blue and are also expressed in units of gpm. Anomalies are calculated from
the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). ............140
Figure A.4 March geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four
years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology values are
shown as grey contours with a contour interval of 30 geopotential metres (gpm). Positive
XV

geopotential height anomalies are shown in yellow and orange, and negative anomalies are
shown in shades of blue and are also expressed in units of gpm. Anomalies are calculated from
the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). ............141
Figure A.5 April geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four
years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology values are
shown as grey contours with a contour interval of 30 geopotential metres (gpm). Positive
geopotential height anomalies are shown in yellow and orange, and negative anomalies are
shown in shades of blue and are also expressed in units of gpm. Anomalies are calculated from
the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). ............142
Figure A.6 May geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four
years of interest, mapped on a 2.5° grid. Monthly geopotential height climatology values are
shown as grey contours with a contour interval of 30 geopotential metres (gpm). Positive
geopotential height anomalies are shown in yellow and orange, and negative anomalies are
shown in shades of blue and are also expressed in units of gpm. Anomalies are calculated from
the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.). ............143
Figure A.7 Contour maps of monthly (December, January, and February) average sea level
pressure anomalies for the four water years of interest, with respect to the 1991 - 2020 reference
period. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset and mapped in units
of hectopascals (hPa), with a contour interval of 2 hPa. Blue lines represent 0 hPa. Original maps
were modified to highlight positive (yellow) and negative (red) anomalies (original image source:
IRI, n.d.). .................................................................................................................................144
Figure A.8 Contour maps of monthly (March, April, and May) average sea level pressure
anomalies for the four years of interest, with respect to the 1991 - 2020 reference period.
Anomalies are calculated from the NCEP-NCAR Reanalysis dataset and mapped in units of
hectopascals (hPa), with a contour interval of 2 hPa. Blue lines represent 0 hPa. Original maps
were modified to highlight positive (yellow) and negative (red) anomalies (original image source:
IRI, n.d.) ..................................................................................................................................145

XVI

List of Abbreviations
AHCCD
CHSS
CoCoRaHS
ECCC
ERA5
ERA5-Land
IRI
KGE
ME
NB
NBDELG
NCEP-NCAR Reanalysis

NWS
QC
RMSE
SAJESS
SJR
SJRB
SWE
USGS
WSC

Adjusted and Homogenized Canadian Climate Data stations
Canadian Historical Snow Survey
Community Collaborative Rain, Hail and Snow Network
Environment and Climate Change Canada
The fifth generation European Centre for Medium-Range Weather
Forecasts (ECMWF) atmospheric reanalysis of the global climate
Reanalysis dataset providing land variables at an enhanced
resolution compared to ERA5
Institute for Climate and Society
Kling-Gupta efficiency
Maine, United States
New Brunswick, Canada
New Brunswick Department of Environment and Local
Government
National Centers for Environmental Prediction (NCEP) and the
National Center for Atmospheric Research (NCAR) Reanalysis
Project
National Weather Service
Quebec, Canada
Root-mean-square error
Saint John River Experiment on Cold Season Storms
Saint John River
Saint John River Basin
Snow water equivalent
United States Geological Survey
Water Survey of Canada

XVII

Acknowledgements
This thesis would not have been possible without the generous support and expertise of
my co-supervisors, Stephen Déry and Ronald Stewart; committee members, Julie Thériault and
Peter Jackson; and external examiner, Guillaume Fortin. My sincere thanks to all of you for your
encouragement, patience, and guidance throughout this journey.
Many thanks to my colleagues at the Northern Hydrometeorology Group and the
Université du Québec à Montréal for the advice and support that you have provided me during
the completion of my Master’s. Special thanks to Katja Winger for assistance in developing the
storm tracks, and to Hadleigh Thompson, Nicolas Leroux, and Dominique Boisvert for their
mentorship. Thank you to NB Power for the support they provided. Jacques Doiron provided
input and advice on this project, and the local knowledge he imparted was invaluable. Thanks
also to Doug Beckett for showing enthusiasm in this project and providing insight about the role
of forestry on river hydrology.
Financial support for this project has been provided by Global Water Futures, which is
funded by the Canada First Research Excellence Fund, the Natural Sciences and Engineering
Research Council of Canada, and the University of Northern British Columbia’s Office of Research
and Graduate Programs.
Finally, thank you to my parents and friends for their support and faith in me. Most of all,
thank you to my partner, Geoff, who helped me overcome my imposter syndrome and
encouraged me to see this project through to completion.

XVIII

Chapter 1 Introduction
1.1 Territorial Acknowledgement
I would like to begin by acknowledging that the land on which this research takes place is
the traditional unsurrendered and unceded territory of the Wolastoqiyik Peoples, and the Saint
John River (SJR) is also known by its original Indigenous name, the Wolastoq. For the purpose of
this study, the waterway will be referred to as the SJR.
1.2 Motivation
With a drainage basin of 55,110 km2 and a length of 673 km, the SJR is the longest in the
Maritimes region of Canada (Cunjak & Newbury, 2005). With its headwaters in northern Maine
(ME), the SJR basin (SJRB) traverses provincial and international borders as it drains water from
southeastern Quebec (QC) before flowing southeast through New Brunswick (NB), where it
empties into the Bay of Fundy (Beltaos et al., 2003).
Being the most frequent natural hazard in Canada (Government of Canada [GC], 2019),
the risk of flooding is a serious concern for the flood-prone SJRB, which is home to over half a
million people—many of whom live along the river and its tributaries (Newton & Burrell, 2016;
New Brunswick Department of Environment and Local Government [NBDELG], 2014). In 2008,
NB experienced its worst spring flood in 35 years, affecting 631 properties and exceeding $23
million in damages (NBDELG, n.d.-a). Following the 2008 spring flood, the region experienced
major spring flooding in 2018, with water levels comparable to or exceeding those of 2008 at
certain points along the river (CBC News, 2018), and again in 2019, bringing the total to three
major events in 12 years.
1

The atmospheric processes that occur in the upper SJRB influence the magnitude and
timing of the spring freshet throughout the basin, but especially in the lower basin (Buttle et al.,
2016; Saint John River Experiment on Cold Season Storms, n.d.). Although there have been
studies conducted on flooding in the SJRB, many of these have been focused on ice jam flooding
(e.g., Beltaos, 1999; Beltaos et al., 2003; Beltaos and Prowse, 2001). In response to the lack of
atmospheric studies in the SJRB, the Saint John River Experiment on Cold Season Storms (SAJESS)
was formed and, as part of the larger SAJESS project, this research used multiple perspectives to
classify and compare the last three major spring flood events (in 2008, 2018, and 2019) by their
causative processes. To perform a comprehensive comparison of the recent major spring flood
events, comparisons were made between these major spring flood events and the 2021 season,
in which no major spring flooding occurred.
1.3 Objectives
The main objectives of this research are:
1. To identify the sequence of events—at various spatial and temporal scales—that led to
the 2008, 2018, and 2019 major spring floods in the SJRB.
2. To compare these major flood events to the 2021 season, in which no major spring
flooding occurred.
1.4 Thesis Structure
This thesis is organised as follows: Chapter 2 consists of a literature review, providing
background information, including physiography and climatology, changing climate, atmospheric
circulation and storms, spring flooding, and causative classification of floods. Chapter 3 provides
2

the sources of data that were used for analysis and data comparisons, as well as the means
through which a number of variables were acquired and / or determined. Chapter 4 includes
results from the data analysis and is divided into three main sections, addressing each of the
following three flood classification perspectives: hydroclimatic, hydrologic, and hydrograph.
Chapter 5 provides a discussion and synthesis of the study and includes the chronology, seasonal
patterns, and rain-on-snow events of the flood and non-flood years; comparisons between flood
and non-flood years; and comparisons to previous studies. Lastly, Chapter 6 consists of
concluding remarks, with a summary of the study and suggestions for future work.

3

Chapter 2 Literature Review
2.1 Physiography and Climatology
The drainage area of the SJR is unequally divided between northern ME (35.7%),
southeastern QC (12.5%), and western NB (51.8%) (International Joint Commission, 1975). The
basin is subdivided into three parts: the upper basin, which starts at the headwaters in ME and
runs to Grand Falls, NB; the middle basin, running from Grand Falls to the Mactaquac Dam, NB;
and the lower basin, from the Mactaquac Dam to the estuary at Saint John, NB. These sub-basins,
as well as the reaches and sub-reaches, were delineated and described in a 2011 report published
by the Canadian Rivers Institute (Kidd et al., 2011) (Figure 2.1; Table 2.1). Cunjak and Newbury
(2005) described the upper basin as being “sparsely populated and in relatively pristine
condition” (p. 954), as opposed to the middle (or central) basin, which is disrupted by several
hydroelectric dams. In total, there are over 200 dams or water control structures within the SJRB
(Kidd et al., 2011).
From its headwaters in ME to its estuary in Saint John, NB, the SJR drops 481 m (Figure
2.2) and passes through several ecological zones. The Canadian portion of the basin comprises
seven different ecoregions that are characterised by distinct regional ecological factors, including
physiography, geology, climate, vegetation, soil, water, and fauna (GC, 2021b). Similarly, the
American portion of the basin is divided into five distinct ecological regions (Figure 2.3). While
much of the SJRB is characterised by a humid continental climate (Ahrens et al., 2016; New
Brunswick Department of Natural Resources, 2007), the southern extent of the basin has a
maritime climate due to the influence of the Bay of Fundy and, consequently, the climate within
the basin becomes warmer and wetter when moving from north to south (Kidd et al., 2011).
4

Based on 1981 - 2010 climate normals (the most recent available information from Environment
and Climate Change Canada [ECCC], 2021), the mean annual air temperature and total
precipitation, respectively, in the upper basin (at Edmundston, NB) is 3.6 °C and 1011 mm, the
middle basin (at Woodstock, NB) is 4.8 °C and 1131 mm, and the lower basin (at Saint John, NB)
is 5.2 °C and 1296 mm. Between 20 % to 33 % of the total annual precipitation in NB falls in the
form of snow, with higher snowfall percentages occurring in the colder, northern regions of the
province (Baronetti et al., 2019; El-Jabi et al., 2013). Cunjak and Newbury (2005) provided basinspecific hydrological information, including: mean annual discharge (approximately 1110 m 3 s-1),
mean annual runoff (900 mm in the south and 640 mm in the headwaters), and timing of the
largest annual runoff (April and May).

5

Figure 2.1 Reaches and sub-reaches of the Saint John River, including start and end points
(black dots). The shaded polygons identify the upper (yellow), middle (red), and lower
(green) basins, and the numbered white circles refer to the sub-reaches described in Table
2.1. Political boundaries are shown with a green line (image source: Kidd et al., 2011, figure
2.1, p. 25).

6

Table 2.1 Description of Saint John River reaches and sub-reaches (data source: Kidd et al., 2011,
table 2.1, p. 24).
Reach
Sub-Reach
1 – Headwaters to Grand Falls, NB
1A – Headwaters to upstream of Baker Brook
● Isolated by natural barrier of Grand Falls River
● Little human influence other than
forestry
1B – Baker Brook River to upstream of Green
River
● Captures inputs from multiple
municipalities and industries
1C – Green River to Grand Falls
● Municipalities and main stem’s 1st
reservoir
2 – Grand Falls to Mactaquac Dam
2A – Little River to upstream of Aroostook
River
● Isolated by Grand Falls and Mactaquac
Dam
● Captures inputs from agriculture
2B – Aroostook River to upstream of Tobique
River
● Flow controlled
2C – Tobique River upstream to Beechwood
Dam
● 2nd reservoir
2D – Beechwood to upstream of Becaquimec
Stream
● Flow regulation, food processing,
municipalities (e.g., Florenceville)
2E – Becaquimec Stream (Hartland) to
Mactaquac Dam
● Municipalities and 3rd and largest
reservoir
3 – Mactaquac Dam to Long Reach (Oak Point) 3A – Mactaquac Dam to Jemseg River
● Long Reach represents upper limits of
● Municipalities of Fredericton and
Saint John River estuary as defined by
Oromocto
saltwater intrusion
3B – Jemseg River to Oak Point
● Grand Lake, Canaan-Washademoak
watershed, and Long Reach
4 – Long Reach (Oak Point) to Saint John
4A – Oak Point to Reversing Falls (estuarial)
● Reversing Falls (switches fresh to
Harbour
saltwater and flow direction reversals)
● Saint John River Estuary
4B – Saint John Harbour
7

● Captures inputs from City of Saint John

Figure 2.2 Elevation of the Saint John River basin, including rivers and main stem hydroelectric
dams (data source: United States Geological Survey [USGS], 2013).

8

Figure 2.3 Ecological regions of the Saint John River basin. The seven ecoregions in the Canadian
portion (red shades) include: Appalachians (117), Northern New Brunswick Highlands (118), New
Brunswick Highlands (119), Saint John River Valley (120), Southern New Brunswick Uplands (121),
Maritime Lowlands (122), and Fundy Coast (123). The five biophysical regions in the American
portion (purple shades) include: Boundary Plateau (1), Saint John Uplands (2), Aroostook Hills (3),
Aroostook Lowlands (4), and Eastern Interior (183) (data sources: GC, 2021b; Maine GeoLibrary,
2017).

9

2.2 Changing Climate
The global climate system consists of several interacting components: atmosphere,
hydrosphere, cryosphere, biosphere, and lithosphere. Evidence of a changing climate can be seen
in several of these components, including a warming of the atmosphere and oceans, a rise in sea
level, and a decrease in snow and ice cover (Intergovernmental Panel on Climate Change, 2019).
A changing climate is especially evident in Canada, where increases in mean temperature are
approximately twice the global mean temperature increase (Zhang et al., 2019). This increase in
temperature is nonuniform, varying spatially and seasonally, with northern Canada and winter
being the location and season of greatest change (Zhang et al., 2019). During the 1981 - 2015
period, most of Canada experienced changes to the cryosphere that include—but are not limited
to—decreases in the annual duration of snow cover, decreases in the seasonal snow
accumulation and, in the Maritimes, a decrease in the seasonal maximum snow water equivalent
(SWE) (Derksen et al., 2019).
In addition to the many resources detailing climate change on a global and national scale,
information is available on the effect of climate change at finer spatial scales, narrowing from an
eastern Canada perspective, to provincial, to the SJRB specifically. Research conducted by Zhang
et al. (2001) found an increase in heavy spring rainfall in eastern Canada, which according to
Buttle et al. (2016), “can be hydrologically significant since these rains may cap frozen soils with
ice layers that restrict infiltration or induce rain-on-snow flooding” (p. 23). Buttle et al. (2016)
noted that, in some Canadian regions, there has been a “shift from snowmelt-dominated flooding
to rain-on-snow or rainfall-runoff flooding” (p. 23) because of a warming climate. This observed
shift could be especially alarming to communities within the SJRB, as rain-on-snow events have
10

been noted as a driving force behind major spring flood events (J. Doiron, personal
communication, 10 June 2021). By the year 2100, the average annual air temperature in NB could
rise by as much as 5.2 °C (El-Jabi et al., 2013), with highest increases in winter (Roy and Huard,
2016). As air temperatures rise, more precipitation is expected to fall, including more falling as
rain in winter, and, consequently, average annual discharge is expected to increase (Swansburg
et al., 2004). As a result of climate change, NB is expected to become “warmer, wetter, stormier
and will experience rising sea levels” (NBDELG, n.d.-b). Based on an analysis of climate and
streamflow data for the SJRB, Hare et al. (1997) found that, since 1972, the spring freshet tended
to occur sooner and with increased volume. From analysing runoff and river ice data, Clair et al.
(1997) found that rivers in northern NB showed a weak trend towards fewer days with ice.
Beltaos and Prowse (2001) stated that the processes governing river ice breakup and jamming
are greatly influenced by climate, and further research on midwinter river ice breakup conducted
by Beltaos et al. (2003) suggested that an increased frequency of such events may be an indicator
of climate change within the SJRB.
2.3 Atmospheric Circulations and Storms
A variety of storm types occur, and many affect the SJRB. Since it is situated in the middle
latitudes, the weather in the SJRB is largely determined and characterised by mid-latitude
cyclones, “especially in fall, winter, and spring” (Ahrens et al., 2016, p. 336) (Figure 2.4). Stewart
et al. (1995) summarised the various types of winter storms and the accompanying precipitation
that occur throughout Canada, including extra-tropical cyclones (the most common), blizzards,
mountain-induced storms, lake effect storms, and polar lows. In Atlantic Canada, Clair et al.

11

(1997) identified global atmospheric circulation patterns and ocean currents as major influences
on the regional climate and listed the following meteorological phenomena as most associated
with extreme events: extratropical cyclones (most often winter storms), tropical cyclones
(hurricanes), or summer severe weather (mesoscale convective systems) (p. xxvii). The lowpressure systems that develop in the northern hemisphere occur as cold, dry northern air and
warm moist Caribbean air interact, with intensity peaking in winter (Clair et al., 1997). One source
of major winter storms affecting the SJRB is the cyclogenesis that occurs near Cape Hatteras (see
Hatteras Low in Figure 2.4), which can produce large storms called northeasters / nor-easters.
These storms typically move in a northeastward trajectory and are characterised by high winds
and heavy rain or snow (Ahrens et al., 2016).

Figure 2.4 Typical winter mid-latitude cyclone paths, with the location of the Saint John River
basin identified (red box) (image source: Ahrens et al., 2016, figure 12.5, p. 341).

12

With respect to NB, Fortin and Dubreuil (2020) noted that atmospheric circulation
influences the province through various systems, including those that bring winter storms (e.g.,
low-pressure systems, frontal systems). With the predicted increases in air temperature and
precipitation because of climate change, NB can expect corresponding increases in freezing rain
and ice storms, which can be costly and dangerous, as they pose a serious risk to residents and
infrastructure (Atlantic Coastal Action Program Saint John [ACAPSJ], 2020). In addition, the
predicted increase in the frequency and severity of storms is evident in NB, as the number of
hurricanes and post-tropical storms is expected to increase in the region (ACAPSJ, 2020).
2.4 Spring River Flooding
Flooding, defined by Whitfield (2012) as “a rising and overflowing of a body of water
especially onto normally dry land” (p. 336), is the most frequent and costly natural disaster in
Canada (GC, 2019; Sandink et al., 2010) and can occur directly (e.g., surface accumulation of
rainfall) or indirectly (e.g., an overflowing river) (Whitfield, 2012). The cause of flooding is
complex, often involving a combination of flood generating mechanisms such as snowmelt
runoff, rain, rain-on-snow, and ice jams (Buttle et al., 2016; Whitfield, 2012). In the SJRB, peak
flows typically occur in early May (Buttle et al., 2016), and the magnitude and timing are highly
dependent on the SWE contained in the snowpack (Buttle et al., 2016; Newton & Burrell, 2016).
The complexity of flood characteristics in the lower SJRB is exacerbated due to the Reversing Falls
at the mouth of the river, which can reverse flow due to tidal influences and cause water to
become increasingly backed up (Inland Waters Directorate, 1974).

13

The SJRB has a long, well documented history of flooding that dates back to 1696. The
1696 spring flood heavily impacted Jemseg, NB, as a “late and very high freshet caused late
planting and crop failures” and made residents consider abandoning the settlement (NBDELG,
n.d.-a). A Flood History Database compiled by the NBDELG is accessible through the “Flooding in
New Brunswick” webpage (NBDELG, n.d.-c), along with links to other flood related resources,
including flood maps, flood photographs, and NB’s Flood Risk Reduction Strategy (NBDELG,
2014). In recent history, the SJRB has been subjected to three major spring flood events which
occurred in 2008, 2018, and 2019 (Figure 2.5).
Newton and Burrell (2016) documented the chronology, causes, and costs of the 2008 SJR
spring flood, declaring it the worst in 35 years, having affected communities along the entire
length of the river. Although the 2008 water levels did not exceed those of 1973, it did exceed it
in duration, lasting a total of 16 days as opposed to 12 (ECCC, 2017). ECCC (2017) described the
conditions prior to the spring freshet as being the “perfect storm,” with the basin having “a
record snowfall; a deep, growing and lasting snow cover well into spring; a delayed peak runoff;
sudden warming; and a forecast of copious amounts of rain.” The flood devastated the region,
affecting over 600 properties and leading to approximately CAD $23 million in damages, with the
most damage having occurred in northwestern NB and downstream of Fredericton (Maugerville
to Jemseg) (Newton & Burrell, 2016). For Edmundston, NB in the upper basin, the 2008 flood is
the benchmark for extreme flooding (J. Doiron, personal communication, 10 June 2021).
Following the 2008 flood, back-to-back major spring floods occurred in 2018 and 2019. Regarding
the 2018 flood, CBC News (2018) reported that “water levels along parts of the river were
comparable to or exceeded those measured during the spring flood of 2008.” Jasmin Boisvert, a
14

Specialist in the Water Sciences section of the NBDELG, described the 2018 and 2019 floods in a
webinar, stating that they were highly dependent on day-to-day weather, and that they mainly
affected the lower SJRB (Jemseg Grand Lake Watershed Association [JGLWA], 2021). Each of
these three major spring floods made the list of Canada’s top 10 weather stories of their
respective years (ECCC, 2017, 2019, 2020).

Figure 2.5 Daily water level above the local datum at the Fredericton hydrometric station for the
2008, 2018, and 2019 water years, with the dashed red line indicating the flood stage (data source:
Water Survey of Canada [WSC], 2021).

2.5 Causative Classification of Floods
As mentioned in Section 2.4, the cause of flooding is complex and often involves a
combination of flood generating mechanisms (Buttle et al., 2016; Whitfield, 2012), which
determine the “time of occurrence, duration, extent, and severity” of flood events (Tarasova et
al., 2019, p. 2). In their review of existing causative classifications of flood events, Tarasova et al.
(2019) explained that there are many different flood classification frameworks, and that these
frameworks typically adopt one of the following three perspectives: hydroclimatic, hydrological
(hydrometeorology and catchment state), and hydrograph. These perspectives occur at varying
spatial and temporal scales, with hydroclimatic being the largest (focusing on large-scale
15

circulation patterns and atmospheric state); then hydrological (catchment scale precipitation
patterns and antecedent catchment state); and finally, the hydrograph-based perspective, which
operates at the finest spatial scale (Tarasova et al., 2019). Many studies choose to approach flood
classification from multiple perspectives, incorporating aspects from each of the hydroclimatic,
hydrological, and hydrograph perspectives (e.g., Brimelow et al., 2015; House & Hirschboeck,
1997; Newton & Burrell, 2016).
2.5.1 Hydroclimatic Perspective
Shelton (2009) defined hydroclimatology as being the intersection of climatology and
hydrology, which includes “energy and moisture exchanges between the atmosphere and the
Earth’s surface and energy and moisture transport by the atmosphere” (p. 7). Adopting a large
synoptic domain, the hydroclimatic perspective is typically focused on linking a specific flood
event to atmospheric pressure and circulation (Shelton, 2009), and often excludes the catchment
state (e.g., soil moisture, snow depth) (Tarasova et al., 2019). Examples of large-scale climatic
activities that affect flooding include seasonal accumulation and melting of snow, anomalies in
upper-level circulation configurations and sea surface temperatures, and decadal-scale
circulation episodes (Hirschboeck, 1988). To distinguish between hydroclimatology and
hydrometeorology, Hirschboeck (1988) explained that flood hydroclimatology “has as its
foundation the detailed focus of hydrometeorological-scale atmospheric activity, while at the
same time seeking to place this activity within a broader spatial and temporal, ‘climatic’
perspective” (p. 30). An example of research conducted from a hydroclimatic perspective within
the SJRB includes that of Beltaos (1999), who conducted a hydroclimatic analysis using flow and
weather data to determine if certain occurrences (winter ice breakup events and high spring flow
16

events) were random or the result of long-term trends. Based on their results, Beltaos (1999)
found that an increase in winter air temperature had effectively increased winter and spring
rainfall amounts, leading to the increased peak flow and corresponding winter ice breakup
events. Regarding spring flood events in the SJRB, most of the information appears to be framed
within a hydrological and / or hydrograph perspective rather than a hydroclimatic one (e.g.,
JGLWA, 2021; NBDELG, n.d.-a; Newton & Burrell, 2016).
2.5.2 Hydrologic Perspective
To analyse the causative classifications of flood events, the hydrological perspective
adopts a catchment to regional spatial scale and combines the following three approaches:
hydrometeorological variables (e.g., precipitation, temperature), the catchment state (e.g., SWE,
soil moisture), and hydrological processes (e.g., infiltration or saturation excess) (Tarasova et al.,
2019). Beginning with season-based classifications, this perspective has since evolved to
encompass “a wide range of complex multi-criteria classifications shaped by local and regional
conditions” to identify flood generating mechanisms at the catchment scale (Tarasova et al.,
2019, p. 6).
Based on a review of available literature and the NBDELG flood database, as well as
information provided by local professionals (e.g., J. Doiron, personal communication, 10 June
2021; JGLWA, 2021), the primary drivers of spring flood generating mechanisms within the SJRB
are day-to-day weather variables. These include temperature, rain, rain-on-snow, and snowmelt
and they would be classified under the hydrological perspective. The previously mentioned study
conducted by Newton and Burrell (2016) that identified the causes, assessment, and damages of
the 2008 flood included a view of the event from a hydrological perspective, with an analysis of
17

the following variables: measured SWE of the snowpack, daily mean air temperatures, and daily
rainfall amounts.
2.5.3 Hydrograph Perspective
The hydrograph-based perspective classifies flood generating mechanisms by assuming
that distinct mechanisms produce different hydrograph patterns (Tarasova et al., 2019).
Hirschboeck (1988) analysed hydrograph characteristics and found that flash flood events
produced by microscale and mesoscale atmospheric activity had steeply sloped hydrographs,
whereas longer duration floods produced by synoptic-scale events had gently sloped
hydrographs. An example of flood classification from a hydrograph-based perspective is the study
conducted by Elliott et al. (1982) which, through an examination of hydrographs, distinguished
between snowmelt- and rainfall-generated peak flow events. As part of their analysis of the 2008
flood, Newton and Burrell (2016) incorporated hydrographs, comparing the daily discharges from
various hydrometric stations along the SJR. During a webinar, Water Sciences Specialist Jasmin
Boisvert overlaid hydrographs (at the Fredericton station) from the 2018 and 2019 spring flood
events for comparison, noting that the 2019 flood exhibited slightly gentler slopes due to a more
gradual melting of snow (JGLWA, 2021). Rather than relying solely on a hydrograph-based
perspective, studies often use hydrographs to supplement other hydroclimatic and / or
hydrological data (e.g., Hirschboeck, 1987; Hirschboeck, 1988; JGLWA, 2021; Newton & Burrell,
2016).

18

Chapter 3 Data and Methods
This chapter provides the sources of data that were used for analysis (Sections 3.1 to 3.3)
and data comparisons (Section 3.4), as well as the means through which a number of variables
were acquired and / or determined (Section 3.5).
3.1 Reanalysis Data
3.1.1 ERA5 and ERA5-Land
The fifth generation ECMWF atmospheric reanalysis of the global climate (ERA5) and the
ECMWF reanalysis dataset providing land variables at an enhanced resolution compared to ERA
(ERA5-Land) are gridded reanalysis datasets that combine model data with observations to form
complete and consistent hourly datasets from 1950 to present. The ERA5 data have a horizontal
resolution of 0.25° (approximately 30 km) and, from a hydroclimatic perspective, were used to
identify the large-scale atmospheric processes that led to the recent major spring flood events.
The main variables of interest for this study include geopotential height (m) at 500 hPa, mean
sea level pressure (hPa), and relative vorticity (s-1) at 850 hPa (Hersbach et al., 2019; Hersbach et
al., 2020). The ERA5-Land data have a horizontal resolution of 0.1° (approximately 9 km) and
were used to identify flood generating mechanisms from a hydrological perspective. The main
variables of interest include runoff (m), snow depth (m), snowfall (m of water equivalent),
snowmelt (m of water equivalent), surface air temperature (K), SWE (m), total precipitation (m),
and volumetric soil water layers 1 and 2 (m 3 m-3) (Muñoz-Sabater, 2019; Muñoz-Sabater et al.,
2021). The periods over which the ERA5 and ERA5-Land data were used are listed in Table 3.1.
Anomalies were calculated relative to the 1991 - 2020 reference period.
19

3.1.2 NCEP-NCAR Reanalysis
Maps of monthly geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) and
monthly sea level pressure anomalies are available through the Columbia Climate School
International Research Institute for Climate and Society (IRI), which is part of The Earth Institute
at Columbia University. The IRI analyses monthly anomalies of geopotential heights using the
National Centers for Environmental Prediction (NCEP) and the National Center for Atmospheric
Research (NCAR) Reanalysis Project (NCEP-NCAR Reanalysis). The anomalies are mapped on a
2.5° latitude / longitude grid and are freely available for download (IRI, n.d.). The periods over
which the NCEP-NCAR Reanalysis data were used are listed in Table 3.1 and anomalies were
calculated relative to the 1991 - 2020 reference period. For this study, maps of monthly
standardised geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) and monthly
standardised sea level pressure anomalies were downloaded for the four years of interest.

20

ERA5-Land

ERA5

Source

4.2.6
3.4
3.4
4.2.2

runoff (m)
snow depth (m)

snow water equivalent (m)

surface air temperature (K)

snowmelt (m of water equivalent)

snowfall (m of water equivalent)

4.1.4

relative vorticity (s-1) at 850 hPa

4.2.4
4.2.5
4.2.4
3.4
4.2.1
4.2.2

4.2.2

4.2.4
4.2.1

4.1.4

4.1.3

Section

mean sea level pressure (hPa)

geopotential height (m) at 500 hPa

Variable

Time Period
30 Apr - 4 May 2008
30 Apr - 7 May 2018
23 - 26 Apr 2019
3 - 11 Apr 2021
Dec 2007, 2017, 2018, 2020
Jan - May 2008, 2018, 2019, 2021
Dec 2007, 2017, 2018, 2020
Jan - May 2008, 2018, 2019, 2021
1 April - 15 May 2008, 2018, 2019, 2021
Oct - May 2011 - 2021
Jan - Apr 1974 - 2016
Oct - Dec 2007, 2017, 2018, 2020
Jan - Sep 2008, 2018, 2019, 2021
1 April - 15 May 2008, 2018, 2019, 2021
1991 - 2020
Oct - Dec 2007, 2017, 2018, 2020
Jan - Sep 2008, 2018, 2019, 2021
1 April - 15 May 2008, 2018, 2019, 2021
1991 - 2021
1 April - 15 May 2008, 2018, 2019, 2021
2011 - 2021
1991 - 2020
Oct - Dec 2007, 2017, 2018, 2020
21

Table 3.1 Time periods of reanalysis data, including the variables used and the corresponding results sections in which they appear.
As applicable, anomalies of variables were calculated relative to the 1991 - 2020 period for all reanalysis datasets.

NCEP-NCAR
Reanalysis

volumetric soil water layers 1 and 2 (m3 m-3)
geopotential height (m) anomalies at 250
hPa, 500 hPa, and 925 hPa
monthly sea level pressure anomalies (hPa)

total precipitation (m)

Dec - May 1991 - 2021
Dec - May 1991 - 2021

4.1.1
4.1.2

4.2.3
4.2.4
4.2.5
4.2.6

4.2.2

4.2.3
3.4
4.2.1

Jan - Sep 2008, 2018, 2019, 2021
Dec - May 1991 - 2021
2011 - 2021
1991 - 2020
Oct - Dec 2007, 2017, 2018, 2020
Jan - Sep 2008, 2018, 2019, 2021
Dec - May 1991 - 2021
1 April - 15 May 2008, 2018, 2019, 2021
Sep - May 1991 - 2021
1 April - 15 May 2008, 2018, 2019, 2021

22

3.2 Weather Station Data
Climate data were acquired from several weather stations. These consist of four stations
operated by ECCC from 2011 to 2021 (GC, 2021a): three in NB (Aroostook, Fredericton, and Oak
Point) and one in QC (St. Camille); three stations in NB operated by the Community Collaborative
Rain, Hail and Snow Network (CoCoRaHS) from 2015 to 2021 (Fredericton 4.0 SSE, Sussex Corner
1.4 NE - KWRC, and Keswick Ridge 2.3 SSE); and two stations in ME operated by the National
Weather Service (NWS) from 2011 to 2021 (Figure 3.1; Table 3.2). Datasets consist of daily values
for the main variables of interest, including air temperature (°C), snow on the ground (cm), and
total precipitation (total rainfall and the water equivalent of the total snowfall in mm).

23

Figure 3.1 Location of weather stations and snow survey sites that were used to assess the
accuracy of the ERA5-Land product within the Saint John River basin, with symbols identifying
the respective collecting agencies, including the Community Collaborative Rain, Hail and Snow
Network (CoCoRaHS) (CoCoRaHS, n.d.), Environment and Climate Change Canada (ECCC) (GC,
2021a), the National Weather Service (NWS) (NWS, n.d.), and the Canadian Historical Snow
Survey (CHSS) (GC, 2021c). Weather station details are provided in Table 3.2.

24

Table 3.2 Weather station metadata (data sources: CoCoRaHS, n.d.; GC, 2021a; NWS, n.d.).
Collecting
Agency

ID

Latitude
(°N)

ECCC1

8100300

46.71

67.72

80.00

1929 - 2022

Fredericton 4.0 SSE

CoCoRaHS CAN-NB-1

45.93

66.63

68.58

2013 - 2022

Sussex Corner 1.4
NE - KWRC

CoCoRaHS CAN-NB-28

45.72

65.46

58.22

2014 - 2022

Keswick Ridge 2.3
SSE

CoCoRaHS CAN-NB-8

45.98

66.87

112.47

2014 - 2022

Station Name
Aroostook, NB

Longitude Elevation
(°W)
(m)

Period of
Record

Caribou, ME

NWS

USW00014
607

46.87

68.02

191.11

1939 - 2022

Fort Kent, ME

NWS

USC00172
878

47.24

68.61

185.93

1893 - 2022

Fredericton, NB

ECCC

8101605

45.92

66.61

35.10

2000 - 2022

Oak Point, NB

ECCC

8103780

45.51

66.10

11.00

2011 - 2022

St. Camille, QC

ECCC1

7056930

46.48

70.22

396.00

1963 - 2022

1

Adjusted and Homogenized Canadian Climate Data (AHCCD) stations.

3.3 Hydrometric Data
To analyse the spring flood events from a hydrograph perspective, hydrometric data were
acquired from 26 locations within the SJRB, including the main stem and its tributaries. The
hydrometric stations located in ME are operated by the United States Geological Survey (USGS),
while the Canadian stations are operated by the Water Survey of Canada (WSC). Each location
provided daily records of water level (m) or flow (m3 s-1) (Figure 3.2; Table 3.3).

25

Figure 3.2 Location of hydrometric data sources, indicating water level (m) or flow (m 3 s-1) (data
sources: USGS, 2021; WSC, 2021). Locations of main stem hydroelectric generating stations are
also identified. Hydrometric station details are provided in Table 3.2.

26

USGS
USGS
WSC
WSC
USGS
WSC
USGS
WSC
WSC
WSC
WSC
WSC

Aroostook River at Washburn, ME

Aroostook River near Masardis, ME

Aroostook River Near Tinker, NB

Becaguimec Stream at Coldstream, NB

Big Black River near Depot Mtn, ME

Big Presque Isle Stream at Tracey Mills, NB

Fish River near Fort Kent, ME

Grande Riviere at Violette Bridge, NB

Iroquois River at Moulin Morneault, NB

Kennebecasis River at Apohaqui, NB

Meduxnekeag River Near Belleville, NB

Nackawic Stream Near Temperance Vale,
NB

Site Name

Collecting
Agency

01AK007

01AJ003

01AP004

01AF009

01AF007

01013500

01AJ004

01010070

01AJ010

01AG003

01015800

01017000

ID

46.05

46.22

45.70

47.46

47.25

47.24

46.42

46.89

46.34

46.82

46.52

46.78

Latitude
(°N)

67.24

67.73

65.60

68.36

67.92

68.58

67.74

69.75

67.47

67.75

68.37

68.16

Longitude
(°W)

240

1,210

1,100

182

339

2,261

484

443

350

6,060

2,310

4,284

Gross
Drainage
Area (km2)

1967 - 2022

1967 - 2022

1961 - 2022

1991 - 2022

1977 - 2022

1990 - 2022

1967 - 2022

1987 - 2022

1973 - 2022

1975 - 2022

1957 - 2022

1930 - 2022

Period of
Record

flow

flow

flow

flow

flow

flow

flow

flow

flow

flow

flow

flow

27

Parameter

Table 3.3 Hydrometric station metadata, including the parameters used in analyses (26 stations) (data sources: USGS, 2021; WSC,
2021).

WSC
USGS

WSC

WSC
WSC
WSC
WSC
USGS
WSC
WSC
WSC
WSC
WSC

Nashwaak River at Durham Bridge, NB

Saint John River at Dickey, ME

Saint John River at Edmundston, NB

Saint John River at Fredericton, NB

Saint John River at Gagetown, NB

Saint John River at Grand Falls, NB

Saint John River at Maugerville, NB

Saint John River at Ninemile Bridge, ME

Saint John River at Oak Point, NB

Saint John River at Saint John, NB

Saint John River at Upper Gagetown, NB

Shogomoc Stream Near Trans Canada
Highway, NB

St. Francis River at Outlet of Glasier Lake,
NB

01AD003

01AK001

01AO011

01AP005

01AP003

01010000

01AO002

01AF002

01AO012

01AK003

01AD004

01010500

01AL002

47.21

45.94

45.85

45.27

45.52

46.70

45.87

47.04

45.77

45.97

47.36

47.11

46.13

68.96

67.32

66.24

66.09

66.08

69.72

66.45

67.74

66.14

66.65

68.33

69.09

66.61

1,350

234

N/A

N/A

N/A

3,473

N/A

21,900

N/A

N/A

15,500

6,941

1,450

1951 - 2022

1918 - 2022

1994 - 2022

1966 - 2022

1923 - 2022

1950 - 2022

1965 - 2022

1930 - 2022

1994 - 2022

1929 - 2022

2001 - 2022

1980 - 1995

1987 - 2022

1962 - 2022

flow

flow

level

level

level

flow

level

flow

level

level

level

flow

flow

28

Williams Brook at Phair, ME

USGS

01017550

46.63

67.95

10

1999 - 2022

flow

29

3.4 Data Comparison
An accuracy assessment of the ERA5-Land product within the SJRB was conducted with
in-situ weather station and snow survey data. The key variables assessed include air temperature,
precipitation, snow depth, and SWE. Data from nine weather stations were used for comparison
with ERA5-Land variables, including four ECCC stations (two of which were AHCCD stations) (GC,
2021a), three CoCoRaHS stations (CoCoRaHS, n.d.), and two NWS stations (NWS, n.d.) (Table 3.2;
Figure 3.1). These nine stations were located throughout the SJRB, in NB, QC, and ME. In addition,
ERA5-Land SWE data were compared to data from three Canadian Historical Snow Survey (CHSS)
sites located across NB (GC, 2021c) (Figure 3.1). The times series of ERA5-Land data were
extracted at the grid cells corresponding to the observation locations.
Datasets were compared by calculating the average root-mean-square error (RMSE),
average bias, and Kling-Gupta efficiency (KGE) (Gupta et al., 2009). From daily data, monthly
means were calculated, and errors were computed from these monthly means using the “rmse”
and “bias” functions of the Metrics library (version 0.1.4) in R. RMSE measures how far predicted
values (ERA5-Land) are from observed values, with lower values of RMSE indicating a better fit.
The bias computes the average amount by which observed values are greater than predicted
values; if a model is unbiased, it should be close to zero. To measure the goodness-of-fit, the KGE
was computed using the “KGE” function of the hydroGOF (version 0.3-2) library in R. KGE values
range from -Inf to 1, with 1 being the most accurate model. Following the methods of Mai et al.
(2022), the KGE values were categorised into corresponding performance levels (Table 3.4). The
time periods over which data were compared was 2011 to 2021 for ECCC and NWS weather
stations, 2015 to 2021 for CoCoRaHS stations, and 1974 to 2016 for CHSS sites.
30

Of all the ERA5-Land variables that were assessed, air temperature and SWE were the
most and least closely aligned to observations, respectively. The close alignment of air
temperature was evident in the overlapping values when plotted, as well as the excellent model
performances at all stations (Figure 3.3; Table 3.5). Alignment of precipitation, snow depth, and
SWE⁠ fluctuated more, which was evident when datasets were compared visually and statistically.
Of the stations considered for precipitation (Table 3.6) and snow depth (Tables 3.7 and 3.8)
comparisons, 50% showed medium or better performances and 50% showed low performances.
SWE was compared at three locations, yielding two low performances and one medium
performance (Table 3.9). Other, larger studies have been conducted to assess the accuracy of
ERA5 on precipitation and temperature over North America (Tarek et al., 2020) and SWE over the
Great Lakes region (Mai et al., 2022). Tarek et al. (2020) found the precipitation and temperature
variables to be “as good as observations over most of North America” (p. 2541), including the
SJRB region. Similarly, Mai et al. (2022) found the SWE variable to be a robust product in the
Great Lakes region. With the temperature variable yielding excellent performances at all stations
and the remaining variables yielding medium or better performances at approximately half of
the stations assessed, the ERA5-Land data were considered fit for use in this study⁠—a decision
bolstered by the positive performances of the reanalysis product in previous studies.

31

Table 3.4 Kling-Gupta efficiency (KGE) values and their corresponding model performance
classifications (Mai et al., 2022).
KGE values

Performance

< 0.48

low

0.48 - 0.65

medium

0.66 - 0.83

good

0.84 - 1.00

excellent

Figure 3.3 Comparison of mean monthly air temperature (°C) between ERA5-Land (black lines)
and weather stations operated by ECCC (red lines) and NWS (blue lines), over a reference period
of 2011 - 2021.

32

Table 3.5 Comparison of mean monthly air temperature (°C) between ERA5-Land and weather
station data, including average bias, average RMSE, and Kling-Gupta efficiency (KGE), over a
reference period of 2011 - 2021.
Fredericton

St Camille

Caribou

Oak Point

Aroostook

Fort Kent

Bias (°C)

-0.14

0.08

0.61

0.08

0.48

-0.65

RMSE (°C)

0.42

0.55

0.75

0.26

0.74

0.85

KGE

0.98

0.98

0.86

0.99

0.90

0.84

Figure 3.4 Comparison of mean monthly total precipitation (mm) between ERA5-Land (black
lines) and weather stations operated by ECCC (red lines) and NWS (blue lines), over a reference
period of 2011 - 2021.

33

Table 3.6 Comparison of mean monthly total precipitation (mm) between ERA5-Land and
weather station data, including average bias, average RMSE, and Kling-Gupta efficiency (KGE),
over a reference period of 2011 - 2021.
Fredericton

St Camille

Caribou

Oak Point

Aroostook

Fort Kent

Bias (mm)

-7.68

-3.54

-5.39

14.48

-4.81

-4.89

RMSE (mm)

11.36

17.00

8.34

18.18

11.22

9.86

KGE

0.18

0.62

0.73

0.33

0.80

0.44

Figure 3.5 Comparison of mean monthly snow depth (cm) from October to May between ERA5Land (black lines) and weather stations operated by ECCC (red lines) and NWS (blue lines), over
a reference period of 2011 - 2021.

34

Table 3.7 Comparison of mean monthly (October to May) snow depth (cm) between ERA5-Land
and weather station data, including average bias, average RMSE, and Kling-Gupta efficiency
(KGE), over a reference period of 2011 - 2021.
Fredericton

St Camille

Caribou

Oak Point

Aroostook

Fort Kent

Bias (cm)

-10.36

-12.06

-6.14

-8.15

-5.26

-15.82

RMSE (cm)

14.71

14.65

7.92

10.34

7.27

19.91

KGE

0.18

0. 62

0.73

0.33

0.80

0.44

Figure 3.6 Comparison of mean monthly (October to May) snow depth (cm) between ERA5-Land
(black lines) and weather stations operated by CoCoRaHS (green lines), over a reference period
of 2015 - 2021.

Table 3.8 Comparison of mean monthly (October to May) snow depth (cm) between ERA5-Land
and weather station data, including average bias, average RMSE, and Kling-Gupta efficiency
(KGE), over a reference period of 2015 - 2021.
CAN-NB-1

CAN-NB-8

CAN-NB-28

Bias (cm)

-2.97

-11.74

-4.41

RMSE (cm)

3.83

15.22

9.66

KGE

0.83

0.30

0.48

35

Figure 3.7 Comparison of mean monthly (January to April) SWE (mm) between ERA5-Land (black
lines) and Canadian historical snow survey data (green lines), over a reference period of 1974 2016.

Table 3.9 Comparison of mean monthly (January to April) SWE (mm) between ERA5-Land and
Canadian historical snow survey data, including average bias, average RMSE, and Kling-Gupta
efficiency (KGE), over a reference period of 1974 - 2016.
Royal Road

Becaguimic

Beechwood

Bias (mm)

43.45

19.91

-2.14

RMSE (mm)

43.97

23.35

25.62

KGE

0.34

0.58

0.39

3.5 Analysis
3.5.1 Hydroclimatic Perspective
A. Monthly Standardised Geopotential Height and Sea Level Pressure Anomalies
From the IRI database, monthly (December to May) standardised geopotential height
anomalies for the four years of interest (2008, 2018, 2019, and 2021) were downloaded and
compiled into figures (one figure per month) (IRI, n.d.). Based on the IRI product, geopotential

36

height was analysed at the standard pressure levels of 250 hPa, 500 hPa, and 925 hPa, as these
are the levels available. Maps were centred on the SJRB, and the domain had the following
bounding box: 91°W, 26°N, 42°W, 67°N. To enable visual comparisons across the four years of
interest, each monthly figure displayed anomaly maps for the four years of interest at each of
the three pressure levels (12 maps per figure). Anomalies were calculated from the NCEP-NCAR
Reanalysis dataset using the 1991 - 2020 reference period and mapped on a 2.5° grid.
Like the geopotential height anomaly maps, monthly average sea level pressure
standardised anomaly maps were also downloaded from the IRI database (IRI, n.d.). These maps
consisted of monthly (December to May) average sea level pressure standardised anomalies for
the four years of interest, with respect to the 1991 - 2020 reference period. With these maps,
two figures were created⁠—one for winter months (December, January, and February) and one
for spring months (March, April, and May)⁠—with each figure displaying monthly sea level
pressure standardised anomalies during the four years of interest.

B. Daily Mean Geopotential Height at 500 hPa
To analyse atmospheric activity during flood events (e.g., Hirschboeck, 1988), daily mean
geopotential heights at 500 hPa were mapped over the following domain: 91°W, 26°N, 42°W,
67°N. These maps were produced with R (version 4.1.0)—with the ‘ggplot2’ and ‘metR’ libraries,
among others—using the ERA5 dataset. For each of the four years of interest, the daily maps that
were produced cover the evolution of the peak spring flow or water level from the upper basin
(at Nine Mile Bridge, ME) to the estuary (at Saint John, NB). From these maps, figures were

37

produced—one per year of interest—to display daily mean geopotential height at 500 hPa during
peak spring flow or water level.

C. Storm Tracks
Using a storm tracking algorithm created by Katja Winger at UQAM (initially created in
2007), low pressure system tracks that correspond with the movement of the low pressure
centres were generated using hourly ERA5 mean sea level pressure (hPa) and 850 hPa relative
vorticity (s-1) data. Other studies that have used versions of this algorithm include Chartrand and
Pausata (2020) and Chen et al. (2022). Using this algorithm, low pressure system trajectories are
created through linear projection, with matches between low pressure centres attempted at
each time step. If no low pressure centres are found within 400 km of a predicted position, the
low pressure system trajectory is completed. A low pressure centre that is not matched to any
previous low pressure system tracks serves as the first centre for a new storm track. For this
study, tracks were generated from December to May of the four years of interest (2008, 2018,
2019, and 2021), and tracks that passed within 500 km of the SJRB were included. The resulting
text files listed storms with hourly points that included latitude, longitude, and sea level pressure
(hPa). With these text files, storm track shapefiles were created and mapped in QGIS (version
3.10.7-A Coruña). The maps were then compiled into two figures—one for winter months
(December, January, and February) and one for spring months (March, April, and May)⁠—with
each figure displaying storm tracks during the four years of interest.

38

3.5.2 Hydrologic Perspective
A. Precipitation and Temperature
Using ERA5-Land data, monthly climate normals (relative to the 1991 - 2020 reference
period) were calculated for surface air temperature (daily mean) and precipitation (total solid
and liquid). To derive liquid precipitation, total snowfall was subtracted from total precipitation.
These climate normals were calculated and visualised (in R version 4.1.0) for the full, upper,
middle, and lower SJRB.
Monthly temperature and precipitation (total, liquid, and solid) anomalies were
calculated and visualised for the four water years of interest (October to September 2008, 2018,
2019, and 2021). These anomalies were calculated for the full, upper, middle, and lower SJRB.
Following Déry and Wood (2005), standardised anomalies were derived by the process of
calculating the difference between the current period and the reference period (1991 – 2020),
and then scaled by dividing by the reference period standard deviation. These standardised
anomalies were used to identify winter (December, January, February) and spring (March, April,
May) air temperature and precipitation anomalies. Seasonal air temperature and precipitation
data were visualised on quadrant plots to identify seasons that were dry-warm, dry-cold, wetwarm, or wet-cold. This analysis was conducted using modified R code from Royé (2020).

B. Heavy Precipitation Events
Heavy precipitation events influence the magnitude and timing of spring flood events in
the SJRB, with heavy snowfall increasing the SWE and heavy rainfall enhancing snowmelt and
increasing runoff (J. Doiron, personal communication, 10 June 2021; JGLWA, 2021; Newton &
39

Burrell, 2016). To identify heavy precipitation events, mean daily (in Coordinated Universal Time)
precipitation values were calculated for the upper, middle, and lower SJRB using ERA5-Land data.
Daily precipitation events < 0.2 mm were considered “trace” amounts and excluded, and from
the remaining values the largest 10 % of daily precipitation events were identified and considered
to be “heavy” events. Others, such as Mekis and Hogg (1999), have followed this approach. Thus,
heavy precipitation event thresholds were defined by the 90 th percentile values, and these
thresholds were specific to the region (upper, middle, and lower SJRB), season, and phase (total,
liquid, and solid precipitation).
For each of the four years of interest, sub-basin specific plots were generated (in R version
4.1.0) to display: the total number of heavy daily precipitation (total, liquid, and solid) events per
season, and heavy daily precipitation (total, liquid, and solid) anomalies per season. Also
generated were heavy daily precipitation (total, liquid, and solid) anomalies per season over the
full SJRB.

C. Snow Water Equivalent (SWE)
Basin-averaged SWE was calculated for the full, upper, middle, and lower SJRB using
ERA5-Land data, and departures from normal were derived for each of the water years of
interest. Monthly anomalies were calculated and visualised on figures as basin-averaged
departures from normal (relative to the 1991 - 2020 reference period).

40

D. Soil Moisture
The mean daily soil moisture was derived by combining the volumetric soil water layers 1
(0-7 cm) and 2 (7-28 cm) from ERA5-Land, and departures from normal (relative to the 1991 2020 reference period) were derived and graphed for each of the water years of interest.
To identify a potential relationship between mean daily soil water content and total daily
runoff, these variables were plotted together from 1 April to 15 May (timeframe of typical peak
spring water level / flow) for each of the four years of interest. Graphs were generated for the
upper, middle, and lower SJRB.
Spatial distribution maps of soil water content anomalies were generated for winter
(December, January, and February) and spring (March, April, and May) for the four years of
interest. To facilitate comparisons between soil water content and air temperature anomalies
during the lead-up to spring flood events, figures were created that featured spatial distribution
maps (for April 2008, 2018, 2019, and 2021) of soil water content anomalies alongside air
temperature anomalies.
3.5.3 Hydrograph Perspective
A. Flood Hydrograph Patterns
To provide an overview of the evolution of flood events along the SJR main stem, flood
hydrographs were created at 26 locations. Like Blöschl et al. (2013), these hydrograph patterns
were displayed together to illustrate flood propagation along the river. These spatial comparisons
were determined for peak spring river flow or stage (depending on the data availability of each
hydrometric station) of each of the four years of interest. In addition, station-specific temporal
41

comparisons of flood events at five locations along the SJR main stem were made by combined
plots—one plot per hydrometric station—displaying river flow or stage for the four years of
interest. These combined hydrographs allowed for location-specific comparisons of flood timing,
duration, and potential flood generating mechanisms based on hydrograph patterns.

B. Evolution of Peak Water Level
To determine water transport times in the SJR, a range of lagged correlations (0 - 10 days)
was conducted between eight main stem hydrometric gauges—from the upper basin (at Nine
Mile Bridge, ME) to the lower basin (at Oak Point, NB)—and the Saint John, NB gauge (at the SJR
estuary). This analysis followed the approach of Albers et al. (2016). Correlations were conducted
in R (version 4.1.0) using Spearman’s rank correlation coefficient, with daily data from 1991 2020. This correlation analysis was repeated for April to mid-May of the four years of interest to
determine how the water transport times during peak spring flow or stage compare to each
other, and to the long-term average.

42

Chapter 4 Results
This chapter includes results from the data analysis and is divided into three main
sections, addressing each of the following three flood classification perspectives: hydroclimatic
(Section 4.1), hydrologic (Section 4.2), and hydrograph (Section 4.3). Ordered from largest to
smallest temporal and spatial scale, this chapter starts with the hydroclimatic perspective,
including results for monthly standardised geopotential and sea level pressure anomalies; daily
mean geopotential height at 500 hPa; and storm tracks. Next is the hydrological perspective,
which includes climate normals (1991 – 2020); monthly and seasonal anomalies; daily rainfall
versus SWE; heavy precipitation events; and antecedent soil moisture. Lastly, results related to
the hydrograph perspective are described, including flood hydrograph patterns; station-specific
temporal comparisons; and evolution of peak water level or flow.
The 2008 flood (23 April to 2 May) was described as “the worst spring flooding in 35 years
along the entire St. John River” (NBDELG, n.d.-a). According to the Flood History Database, the
rapid water level rise was the result of melting snow and rain, which led to high runoff rates and
flooding along the length of the SJR and its tributaries. The flood causal factors were listed as
record breaking snowfalls in winter, a late spring thaw, heavy rain, and warm weather. Following
the 2008 flood, the next major spring flood occurred in 2018 (27 April to 12 May), with water
levels comparable to or exceeding those of 2008 at certain points along the river (CBC News,
2018). Like 2008, the Flood History Database also listed mild weather, snowmelt, and heavy rain
as causal factors. In addition, 2018 had ice jams and high tides that contributed to flooding. Major
spring flooding occurred again the following year in 2019 (19 April to 6 May) and, like the previous
two major flood events, mild weather, snowmelt, and heavy rain were contributing factors. As in
43

2018, ice jams were an additional contributing factor during the 2019 flood. The 2008 flood
affected the entire SJRB, whereas the lower basin was most impacted during the 2018 and 2019
floods (JGLWA, 2021). Using multiple perspectives, the current study identified and compared
the causal factors that led to the 2008, 2018, and 2019 spring flood events. The results presented
herein confirm the previously known causal factors, while more thoroughly quantifying and
interpreting the spatial and temporal characteristics of these features.
4.1 Hydroclimatic Perspective
4.1.1 Monthly Standardised Geopotential Height Anomaly
Monthly standardised geopotential height anomalies (December to May) at 250 hPa, 500
hPa, and 925 hPa for the four years are mapped in Figures 4.1 to 4.6. Although they differed
spatially, December geopotential height anomalies during the flood years were mostly negative.
In reference to the SJRB, these negative anomalies were located to the northeast in 2008 and to
the east in 2019. In 2018, the 250 and 500 hPa anomalies covered the SJRB and extended
westward, while the 925 hPa anomalies were to the northwest. In contrast to the flood years,
the non-flood year exhibited positive anomalies in December, located to the east of the SJRB
(Figure 4.1). Unlike December, January anomalies were primarily positive during the flood years,
with the non-flood year having exhibited high (> 2.5 standard deviations) positive anomalies
north of the SJRB and negative anomalies to the southeast (Figure 4.2). During the remaining
months, no clear pattern emerged among the flood years, with each year differing with respect
to the presence and location of anomalies, and whether those anomalies were positive or
negative (Figures 4.3 - 4.6). There is no clear distinction between the flood years and the nonflood year—except during December and January. In February, similar patterns were seen in the
44

non-flood year and in 2019, with positive anomalies to the north and south of the SJRB and
negative anomalies to the east. Anomalies from March to May vary greatly across all four years.

45

2008

2018

2019

2021

Figure 4.1 December standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the four water years of interest,
mapped on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 60
geopotential metres (gpm). Positive standardised geopotential height anomalies are shown in yellow and orange, and negative
anomalies are shown in shades of blue, with shading starting at ±1 standard deviation and a contour interval of 0.5 standard deviation.
Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.).
46

925
hPa

500
hPa

250
hPa

Dec

2008

2018

2019

2021

Figure 4.2 January standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the four water years of interest, mapped
on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 60 geopotential
metres (gpm). Positive standardised geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown
in shades of blue, with shading starting at ±1 standard deviation and a contour interval of 0.5 standard deviation. Anomalies are
calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.).
47

925
hPa

500
hPa

250
hPa

Jan

Feb

250
hPa

500
hPa

925
hPa

2008

2018

2019

2021

Figure 4.3 February standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the four water years of interest, mapped
on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 60 geopotential
metres (gpm). Positive standardised geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown
in shades of blue, with shading starting at ±1 standard deviation and a contour interval of 0.5 standard deviation. Anomalies are
calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.).

48

2008

2018

2019

2021

Figure 4.4 March standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the four water years of interest, mapped
on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 60 geopotential
metres (gpm). Positive standardised geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown
in shades of blue, with shading starting at ±1 standard deviation and a contour interval of 0.5 standard deviation. Anomalies are
calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.).
49

925
hPa

500
hPa

250
hPa

Mar

Apr

250
hPa

500
hPa

925
hPa

2008

2018

2019

2021

Figure 4.5 April standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the four water years of interest, mapped
on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 60 geopotential
metres (gpm). Positive standardised geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown
in shades of blue, with shading starting at ±1 standard deviation and a contour interval of 0.5 standard deviation. Anomalies are
calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.).

50

May

250
hPa

500
hPa

925
hPa

2008

2018

2019

2021

Figure 4.6 May standardised geopotential height anomalies (at 250, 500, and 925 hPa) for the four water years of interest, mapped
on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 60 geopotential
metres (gpm). Positive standardised geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown
in shades of blue, with shading starting at ±1 standard deviation and a contour interval of 0.5 standard deviation. Anomalies are
calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference period (IRI, n.d.).

51

4.1.2 Monthly Standardised Sea Level Pressure Anomaly
Gridded maps of monthly standardised sea level pressure anomalies (December to May)
for the four years are shown in Figures 4.7 and 4.8. A visual inspection of the monthly
standardised sea level pressure anomaly maps, which are centred on the SJRB, did not yield any
clear patterns differentiating between the flood years and the non-flood year, as certain months
displayed similar patterns among flood and non-flood years (e.g., February in Figure 4.7) and
dissimilarity among flood years (e.g., April in Figure 4.8).

52

2018

2019

2021

53

Figure 4.7 Monthly (December, January, and February) average sea level pressure standardised anomalies for the four water years of
interest, with respect to the 1991 - 2020 reference period. Standardised anomalies are calculated from the NCEP-NCAR Reanalysis
dataset and mapped in an interval of 0.5 standard deviation (image source: IRI, n.d.).

Feb

Jan

Dec

2008

Mar

Apr

May

2008

2018

2019

2021

Figure 4.8 Monthly (March, April, and May) average sea level pressure standardised anomalies for the four water years of interest,
with respect to the 1991 - 2020 reference period. Standardised anomalies are calculated from the NCEP-NCAR Reanalysis dataset and
mapped in an interval of 0.5 standard deviation (image source: IRI, n.d.).

54

4.1.3 Daily Mean Geopotential Height at 500 hPa
Daily mean geopotential height maps at 500 hPa are displayed in Figures 4.9 to 4.12. For
each of the four years, the daily maps cover the period of peak spring flow or water level from
the upper basin (at Nine Mile Bridge, ME) to the estuary (at Saint John, NB). Using daily 500 hPa
geopotential height maps to analyse atmospheric activity during flood events follows the
approach of Hirschboeck (1988).
On the 2008 daily maps, there were two low pressure systems. One was west of the SJRB
that moved away from the basin on a northwest trajectory; and another was southeast of the
basin over the Atlantic Ocean. In addition to the two low pressure systems, a high pressure
system was situated to the northeast of the basin over the Labrador Sea, which moved southwest
towards the basin until it dissipated on 4 May (Figure 4.9). In comparison to 2008, the contour
lines on the 2018 maps had a stronger gradient. Early in the timeframe, a low pressure system
was situated just south of the basin (30 April - 1 May 2018), with a high pressure system in the
southern portion of the domain (Figure 4.10). During the 2019 period, a low pressure system that
was located just south of the basin dissipated as it moved east (Figure 4.11). Compared to the
flood years, during the non-flood year (2021) the peak spring flow and water level occurred
approximately one month earlier in the season. Hence, the days covered are approximately one
month ahead of the flood years. The non-flood year displayed multiple closed centres during the
peak spring flow period, with a low pressure system that moved southeast across the basin and
settled over the Atlantic Ocean (Figure 4.12).

55

30 April 2008

1 May 2008

3 May 2008

4 May 2008

2 May 2008

Figure 4.9 Daily mean geopotential height in metres at 500 hPa during the spring 2008 peak water
level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 30 April - 4 May (five
days). Maps were produced with R (version 4.1.0) using the ERA5 dataset.

56

30 April 2018

1 May 2018

2 May 2018

3 May 2018

4 May 2018

5 May 2018

6 May 2018

7 May 2018

Figure 4.10 Daily mean geopotential height in metres at 500 hPa during the spring 2018 peak
water level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 30 April 30 - 7
May (eight days). Maps were produced with R (version 4.1.0) using the ERA5 dataset.

57

23 April 2019

24 April 2019

25 April 2019

26 April 2019

Figure 4.11 Daily mean geopotential height in metres at 500 hPa during the spring 2019 peak
water level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 23 April - 26
April (four days). Maps were produced with R (version 4.1.0) using the ERA5 dataset.

58

3 April 2021

4 April 2021

5 April 2021

6 April 2021

7 April 2021

8 April 2021

9 April 2021

10 April 2021

11 April 2021

Figure 4.12 Daily mean geopotential height in metres at 500 hPa during the spring 2021 peak
water level (from Ninemile Bridge in the headwaters to Saint John at the mouth), 3 - 11 April
(nine days). Maps were produced with R (version 4.1.0) using the ERA5 dataset.

59

4.1.4 Storm Tracks
Monthly low pressure system tracks (December to May), which correspond with the
movement of the low pressure centres, for the four years are displayed in Figures 4.13 and 4.14.
Storm tracks were included if they passed within 500 km of the SJRB. Note that none of the storm
tracks passed the boundary between months. In addition, storms that produce heavy rainfall over
the SJRB, as well as the time series of surface pressure evolution, are shown in Figures 5.2 to 5.5.
The most obvious pattern is that they almost all had a SW to NE direction of movement.
This was independent of month and whether a flood or non-flood year was occurring.
The storm tracks did vary as to their proximity and direction relative to the SJRB. Most
storm tracks passed across a north-south longitude through the basin. With a total of 21, 2019
had the most storm tracks that passed south of, over top of, or north of the basin (Table 4.1). The
second highest was 2008, followed by 2018 and the non-flood year (2021), respectively.
Collectively, numerous storm systems affected the SJRB.
Although the directionality was similar, some clear distinctions can be made between the
flood and non-flood years with respect to the number of tracks. The non-flood year had the
fewest total storm tracks, and zero tracks that passed north of the basin. April was the only month
that the non-flood year had more tracks than one of the flood years. In all other cases, the nonflood year either had the fewest tracks (Jan, May) or was tied for the fewest tracks (Dec, Feb,
Mar). However, regarding tracks that passed directly over the basin, the non-flood year tied with
2019 for the second highest number. For tracks that passed south of the basin, the non-flood
year had the second highest number, and for tracks that passed north of the basin, it had the
fewest (zero). When narrowing the focus to spring storm tracks (Mar to May), the outcome is
60

similar: the non-flood year had the fewest total tracks, fewest that passed south of the basin,
tied with 2019 for second highest number of tracks directly over the basin, and tied with 2018
for the fewest number of tracks north of the basin (zero).
It should be noted that some storms did not pass south, over, or north of the SJRB. In
many of these cases, storms were initiated east of the SJRB and would not have directly affected
the SJRB. In other cases, such as December 2021, the storm track was mainly from south to north
but remained west of the SJRB; such a track may have contributed to SJRB precipitation.

Table 4.1 Number of storm tracks that passed south of the Saint John River basin (SJRB), directly
over the SJRB, or north of the SJRB. The two entries separated by a slash are the number from
December (Dec) to May and the number from March (Mar) to May of the four years of interest.
Storm tracks are mapped in Figures 4.13 and 4.14.
Number of Storm Tracks Dec to May / Mar to May
south of SJRB over SJRB north of SJRB
total
2008

9/2

4/3

6/1

19 / 6

2018

3/0

9/4

2/0

14 / 4

2019

9/4

6/3

6/3

21 / 10

2021

4/0

6/3

0/0

10 / 3

61

2018

2019

2021

62

Figure 4.13 Monthly low pressure system tracks for December (Dec), January (Jan), and February (Feb) generated from ERA5 data for
the four water years of interest, with first points identified by the code represented by black dots and the Saint John River basin
delineated by the black polygon.

Feb

Jan

Dec

2008

2018

2019

2021

63

Figure 4.14 Monthly low pressure system tracks for March (Mar), April (Apr), and May generated from ERA5 data for the four water
years of interest, with first points identified by the code represented by black dots and the Saint John River basin delineated by the
black polygon.

May

Apr

Mar

2008

4.2 Hydrologic Perspective
4.2.1 Climate Normals
Climate normals of air temperature and precipitation (liquid and solid phase) for the full,
upper, middle, and lower SJRB are shown in Figure 4.15. When averaged over the full basin, mean
daily air temperature ranges from a minimum of -10.9 °C in January to a maximum of 18.7 °C in
July, while precipitation ranges from a minimum of 75.6 mm in February to a maximum of 107.1
mm in October. Similarly, when each subsection of the basin is averaged separately, January
yields the coldest mean daily air temperature values in the upper, middle, and lower basins
(-12.3 °C, -11.1 °C, and -8.6 °C, respectively), with peak air temperature in July (18.2 °C, 18.8 °C,
and 19.1 °C, respectively). Thus, mean air temperature decreases when moving north from the
lower to upper basin. Like the full basin, the upper and middle basins receive minimum
precipitation values in February (69.4 mm and 74.9 mm, respectively). However, these sub-basins
experience maximum precipitation amounts in July rather than October (114.5 mm and 106.7
mm, respectively). The lower basin is unique in that it experiences minimum precipitation in
August (79.7 mm) and maximum precipitation in December (115.5 mm).
As expected, the amount of precipitation occurring in the solid phase increases when
moving north from the lower to upper basin. The lower basin receives, on average, 24 % of annual
precipitation in the solid phase. The average annual contribution of solid phase precipitation
increases to 27 % in the middle basin and 30 % in the upper basin. For the upper and middle
basins, December to March receive > 50 % of monthly precipitation in the solid phase. In the
lower basin, December receives < 50 % of precipitation in the solid phase. On average, in each
64

section of the basin, some liquid precipitation occurs even in the coldest month (January) ranging
from 44 % in the lower basin to 27 % in the upper basin.

Figure 4.15 Monthly climate normals of air temperature, precipitation, and its phase for the Saint
John River basin (SJRB) using the ERA5-Land dataset, in reference to the 1991 - 2020 period.
Normals are shown for the full, upper, middle, and lower SJRB.

65

4.2.2 Monthly Anomalies (Temperature, Precipitation, SWE, and Soil Water)
For each of the four years of interest (2008, 2018, 2019, and 2021), monthly anomalies of
precipitation (total, liquid, and solid), SWE, air temperature, and soil water content are shown
for each of the full, upper, middle, and lower SJRB in Figures 4.16 to 4.21.
Mean monthly air temperature anomalies varied substantially. They were exceedingly
positive throughout the basin in the non-flood year, with winter and spring months all observing
positive temperature anomalies (Figure 4.16). In contrast, except for July and August, 2019
observed negative temperature anomalies throughout the basin. Anomalies varied during the
remaining two flood years, 2008 and 2018, with both positive and negative anomalies during
winter and spring months.

66

Figure 4.16 Mean monthly air temperature (T) anomalies for the full, upper, middle, and lower Saint John River basins for
the four water years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period.

67

Monthly precipitation anomalies display temporal and spatial variability among flood
years. Basin-wide January total precipitation anomalies are negative in 2008 (like the non-flood
year) and positive in 2018 and 2019 (Figure 4.17). Spatial variability can also be observed in some
months, with February 2018 experiencing negative precipitation anomalies in the upper basin
and positive anomalies elsewhere. One difference between the flood years and the non-flood
year are the December anomalies, with all areas experiencing slightly negative anomalies during
flood years and positive anomalies (nearly 60 % in the lower basin) during the non-flood year.
While mean monthly liquid precipitation anomalies (Figure 4.18) display close similarity to the
total precipitation anomalies, solid precipitation anomalies display some variation (Figure 4.19).
This is especially true for the December solid precipitation anomalies, which in the non-flood year
were negative across the entire basin (with the highest anomaly of -40 % occurring in the lower
basin). Thus, while December total precipitation anomalies were positive in the non-flood year,
this was solely attributable to liquid precipitation rather than solid. For the flood years, 2008 and
2018 experienced positive December solid precipitation anomalies whereas negative anomalies
were observed in 2019.

68

Figure 4.17 Mean monthly total precipitation (TP) anomalies for the full, upper, middle, and lower Saint John River basins for
the four water years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period.
69

Figure 4.18 Mean monthly liquid precipitation (LP) anomalies for the full, upper, middle, and lower Saint John River basins for
the four water years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period.
70

Figure 4.19 Mean monthly solid precipitation (SP) anomalies for the full, upper, middle, and lower Saint John River basins for
the four water years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period.
71

The mean monthly SWE anomalies are more striking. These were negative in the nonflood year in all areas and months—except for October in the upper basin (Figure 4.20). Among
the flood water years, 2008 and 2019 experienced positive SWE anomalies in winter (December,
January, and February) and spring (March, April, and May) months throughout the basin. In
contrast, the 2018 flood water year had negative anomalies throughout the basin in December,
with these negative anomalies continuing through to March in the lower basin.

72

Figure 4.20 Mean monthly snow water equivalent (SWE) anomalies for the full, upper, middle, and lower Saint John River
basins for the four water years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 2020 period.

73

For mean soil water anomalies, patterns are not as evident, with monthly anomalies
differing spatially and temporally (Figure 4.21). During the lead-up to peak spring flow, which
typically occurs in late April to early May, the non-flood year had negative soil water anomalies
in the full, upper, and middle SJRB, with a positive anomaly during May in the lower basin. During
the 2008 flood year, April soil water anomalies were positive everywhere except the upper basin,
whereas May anomalies were positive in the upper basin and negative in the middle and lower
basins. Negative anomalies occurred in April and May of 2018, except for April in the lower basin.
During April and May of 2019, positive soil water anomalies occurred throughout the basin.

74

Figure 4.21 Mean soil water (SW) anomalies for the full, upper, middle, and lower Saint John River basins for the four water
years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period.

75

4.2.3 Seasonal Precipitation and Temperature Anomalies
Standardised precipitation and air temperature anomalies for the full, upper, middle, and
lower SJRB during winter and spring are shown in Figures 4.22 and 4.23, respectively. During the
winter, none of the four years of interest were greater than two standard deviations from the
mean precipitation or air temperature (Figure 4.22). When averaged over the full basin, all four
years were wetter than normal, but differed with respect to air temperature. The non-flood year
fell well within the wet-warm quadrant, whereas the three flood years were either within (2019)
or near (2008 and 2018) the wet-cold quadrant. This pattern is also observed when viewing the
upper, middle, and lower basins separately, with the non-flood year being wet-warm in the
middle and lower basins and dry-warm (slightly drier than normal) in the upper basin. Thus,
throughout the basin, the non-flood year experienced warmer than normal temperatures and
greater than normal precipitation⁠—except for the upper basin, which was slightly drier than
normal. For the three flood years, they remained in or near the wet-cold quadrant (all being
wetter than normal, with some years being slightly warmer than normal).

76

Figure 4.22 Standardised winter (December, January, and February) precipitation and air temperature
anomalies for the full, upper, middle, and lower Saint John River basins using the ERA5-Land dataset,
in reference to the 1991 - 2020 reference period. The four years of interest are identified with white
dots. Years that fall outside of the hashed box have a difference greater than two standard deviations
from the reference period.

77

As with winter, none of the four years of interest was greater than two standard
deviations from the mean precipitation or air temperature (Figure 4.23). In comparison to winter,
the standardised spring precipitation and air temperature anomalies showed more variability
amongst the four years of interest. In each area of the basins, the non-flood year was within the
dry-warm quadrant, except for the lower basin, which was also warmer than normal but did not
experience a precipitation anomaly. When averaged across the full basin, the three flood years
fell within different quadrants: 2008 was dry-cold, 2018 was wet-warm, and 2019 was wet-cold.
When viewing subsections of the basin separately, 2018 and 2019 remained in the wet-warm
and wet-cold quadrants, respectively. However, 2008 deviated in the upper basin where it fell
within the wet-cold quadrant rather than the dry-cold.

78

Figure 4.23 Standardised spring (March, April, and May) precipitation and air temperature anomalies
for the full, upper, middle, and lower Saint John River basins using the ERA5-Land dataset, in reference
to the 1991 - 2020 reference period. The four years of interest are identified with white dots. Years
that fall outside of the hashed box have a difference greater than two standard deviations from the
reference period.

79

4.2.4 Rainfall versus SWE
Flooding is not just about monthly or seasonal values. It can be influenced by shortduration events. To examine this issue, Figure 4.24 shows total rainfall and SWE from 1 April to
15 May for all four years. The overall reduction of SWE was occurring while several significant
precipitation events occurred. These events occurred in all sub-basins, although their magnitudes
were generally greatest in the lower sub-basin.
To identify and analyse the occurrence and magnitude of rain-on-snow events,
cumulative rainfall and snowmelt from 1 April to 15 May for the four years are shown in Figure
4.25. From the ERA5-Land product, cumulative rainfall was derived by subtracting snowfall from
total precipitation. Snowmelt, which is a variable taken directly from ERA5-Land, accounts for
factors such as sublimation and evaporation. During the flood years, cumulative rainfall and
snowmelt were closely aligned, indicating the presence of rain-on-snow events. The effects of
these rain-on-snow events are evident in the sharp decreases in mean daily SWE (Figure 4.24)
and sharp increases to cumulative snowmelt (Figure 4.25). In contrast, during the non-flood year,
cumulative rainfall and snowmelt were not closely aligned, as snowmelt largely preceded the
rainfall season.

80

81

Figure 4.24 Daily total rainfall and mean snow water equivalent (SWE) from 1 April - 15 May for the four years, calculated for the full,
upper, middle, and lower Saint John River basins using the ERA5-Land dataset.

82

Figure 4.25 Cumulative (cum) rainfall and snowmelt from 1 April - 15 May for the four years, calculated for the full, upper, middle, and
lower Saint John River basins using the ERA5-Land dataset.

4.2.5 Heavy Precipitation Events
A critical issue to address is the degree to which precipitation occurred in large
accumulation events, with heavy one day snowfall events in winter increasing the SWE and heavy
rainfall events in spring enhancing snowmelt and increasing runoff. Heavy precipitation events
had a daily (in Coordinated Universal Time) temporal scale. As described in Section 3.5.2.B, the
thresholds⁠—specific to phase, season, and region⁠—that defined heavy precipitation events are
listed in Table 4.2, with thresholds generally decreasing when moving from the lower to upper
basins. For example, the rainfall threshold during spring decreases from 11.1 mm in the lower
basin to 9.5 mm in the upper basin.

Table 4.2 Heavy precipitation event thresholds (mm), defined by the 90 th percentile values,
specific to the phase (total, liquid, and solid precipitation), season (autumn: September, October,
November; winter: December, January, February; spring: March, April, May), and region (upper,
middle, and lower Saint John River basins). Thresholds were calculated from ERA5-Land data
relative to the 1991 - 2020 period.
Saint John River Basin
Variable
Total
Precipitation

Rainfall

Snowfall

Season

Upper

Middle

Lower

Autumn

12.9

14.6

16.0

Winter

9.5

12.2

14.5

Spring

11.0

12.0

12.6

Autumn

12.7

14.1

15.4

Winter

9.6

11.6

12.2

Spring

9.5

10.4

11.1

Autumn

5.5

5.8

5.5

Winter

7.1

7.9

9.5

Spring

6.8

6.8

7.6
83

The number of heavy precipitation events was examined within the four years. When
examining the total number of heavy daily precipitation events, no discernible patterns exist
between the flood and non-flood years, as the number of events varied spatially and temporally.
For example, the number of heavy daily liquid precipitation events in spring varied in the upper
basin between three (in 2008) and nine (in 2019) within the flood years and was five in the nonflood year (Figure 4.26).
Heavy daily precipitation anomalies per season and phase for the SJR sub-basins are
shown in Figure 4.27. For the flood years, anomalies varied spatially and temporally. For example,
in the spring of 2008 there were no heavy daily total precipitation anomalies in the upper and
middle basins, but there was a negative anomaly (of -4 days) in the lower basin. In contrast, the
other two flood years had positive anomalies in all sections of the basin, and the non-flood year
had positive anomalies in the upper and lower basins with no anomaly in the middle basin.
When averaged across the full basin, all flood years had positive winter solid precipitation
anomalies. In contrast, the non-flood year was characterised by a negative winter solid
precipitation anomaly (Figure 4.28).

84

85

Figure 4.26 The total number of heavy daily precipitation (total, liquid, and solid) events per season for the four years of interest,
specific to the upper (top row), middle (middle row), and lower (bottom row) Saint John River basin.

86

Figure 4.27 Heavy daily precipitation (total, liquid, and solid) anomalies per season for the four years of interest, specific to the upper
(top row), middle (middle row), and lower (bottom row) Saint John River basin.

87

Figure 4.28 Heavy daily precipitation (total, liquid, and solid) positive (pos) and negative (neg) anomalies per season for the four years
of interest, over the full Saint John River basin.

4.2.6 Antecedent Soil Moisture
Antecedent soil moisture is another critical issue to address because high soil moisture
content could potentially decrease infiltration, simultaneously increasing runoff and flood
potential. The relationship between total runoff and soil water content across the full, upper,
middle, and lower SJRB for each of the four years are shown in Figure 4.29. As expected, the total
runoff and soil water content tended to peak and fall together, which is especially evident in 2019
in the upper and middle basins. In addition, cumulative runoff is displayed in Figure 4.30 and
shows steep increases in cumulative runoff during flood years. These steep increases in
cumulative runoff generally coincided with the increases in daily total runoff and mean soil water
content. In contrast, cumulative runoff during the non-flood year was steady during the 1 April
to 15 May period, with no significant increases.
Spatial distributions of winter and spring soil water anomalies are shown in Figure 4.31,
with variation of positive and negative anomalies across the four years. No discernible patterns
distinguished the non-flood year from the flood years, as the winter anomalies were similar to
those in 2008 and 2018 (negative anomalies in the upper and middle, with positive anomalies in
the lower), with spring anomalies differing spatially across all four years.
To identify a potential relationship between soil water and air temperature anomalies
during the lead-up to peak spring flow, April spatial distribution maps for the four years of
interest are displayed in Figure 4.32. No clear relationship emerged among the three flood years,
as they differed with respect to both variables. However, in comparison to the three flood years,

88

soil water anomalies observed during the non-flood year were more negative in the middle to
lower basins, and the air temperature anomalies were higher throughout the basin.

89

90

Figure 4.29 Daily total runoff and mean soil water content from 1 April - 15 May for the four years of interest calculated for the full,
upper, middle, and lower Saint John River basins using the ERA5-Land dataset.

Figure 4.30 Cumulative (cum) runoff from 1 April - 15 May for the four years of interest calculated for the full, upper, middle, and
lower Saint John River basins using the ERA5-Land dataset.

91

92

Figure 4.31 Spatial distribution of soil water (SW) (liquid and solid) anomalies for winter (December, January, and February) and
spring (March, April, and May) for the four years of interest. Anomalies were derived from the ERA5-Land dataset in reference to
the 1991 - 2020 period.

93

Figure 4.32 Spatial distribution of soil water (SW) (liquid and solid) (top row) and 2 m air temperature (2m T) (bottom row) anomalies
in April for the four years of interest. Anomalies were derived from the ERA5-Land dataset in reference to the 1991 - 2020 period.

4.3 Hydrograph Perspective
4.3.1 Flood Hydrograph Patterns
To provide an overview of the evolution of peak spring flow or level within the SJRB,
hydrograph patterns for the four years of interest are shown in Figures 4.33 to 4.36. Following
the approach of Blöschl et al. (2013), these figures display hydrographs of water flow or level
throughout the SJRB to illustrate flood propagation.
Although they differ in shape (steepness and number of peaks), the hydrograph patterns
of the three flood years displayed similarity in peak flood timing (late April to early May). In
contrast, the non-flood year peaked in early April and then tapered off steadily—except for some
fluctuation in the upper basin—through to mid-May.
When comparing the three flood years, flood patterns in the upper basin varied greatly
with respect to steepness, peak shape, and number of peaks. For example, the 2018 flood peaked
very quickly (steeply sloped), whereas the flood response in 2008 and 2019 was relatively slower
(less steep). In addition, 2008 and 2019 displayed high fluctuation in peak flow (multiple peaks
and falls) in the upper and middle portions of the basin, whereas 2018 displayed a flatter peak
with less fluctuation; 2018 remained at peak flow and level longer than the other two flood years.
The flood level patterns in the lower basin during 2008 displayed a gradual increase leading to
peak flood level, whereas the 2018 and 2019 floods had steeper slopes, indicating a faster rise in
water level.

94

Figure 4.33 Flood hydrographs for 1 April - 15 May 2008 derived from 26 hydrometric stations
across the Saint John River Basin. The upper, middle, and lower sections of the river are
delineated by the thick black lines.

95

Figure 4.34 Flood hydrographs for 1 April - 15 May 2018 derived from 26 hydrometric stations
across the Saint John River Basin. The upper, middle, and lower sections of the river are
delineated by the thick black lines.

96

Figure 4.35 Flood hydrographs for 1 April - 15 May 2019 derived from 26 hydrometric stations
across the Saint John River Basin. The upper, middle, and lower sections of the river are
delineated by the thick black lines.

97

Figure 4.36 Hydrographs for 1 April - 15 May 2021 derived from 26 hydrometric stations across
the Saint John River Basin. The upper, middle, and lower sections of the river are delineated by
the thick black lines.

98

4.3.2 Station-Specific Temporal Comparisons
To further analyse flood response, station-specific temporal comparisons are displayed in
Figure 4.37. This figure was created by combining hydrographs from all four years, at five
locations along the SJR main stem.
Among the flood years, the 2019 flood wave arrived earliest in the season, followed by
2018 and 2008. In terms of the maximum water flow and level, these values varied spatially and
temporally among flood years. For example, at the headwaters the highest water flow occurred
in 2018 (1087.19 m3 s−1), but at the mouth the highest water level occurred in 2018 (5.68 m). As
previously shown in Section 4.3.1, Figure 4.3.7 also illustrates the flatter, longer sustained peak
of the 2018 flood in comparison to the other two flood years.
When all years are displayed together, it is evident that the non-flood year varied greatly
in comparison to the three flood years. In terms of timing of the peak, the non-flood year peaked
in early April and was more stable as it had fewer and lower peaks. Although the water level in
early April approached the flood stage in the lower basin, it did not exceed it, and water levels
remained well below flood stage during the entire spring season in all sections of the basin.

99

Figure 4.37 Hydrometric station-specific temporal comparisons at five locations along the main stem
of the Saint John River, for the four years of interest. For stations reporting water level, the flood
stage is indicated by a horizontal dashed line. The upper, middle, and lower sections of the river are
delineated by the thick black lines.

100

4.3.3 Lagged Correlations of Hydrometric Gauges
The lag time correlations between eight SJR main stem hydrometric gauges and the Saint
John, NB gauge at the estuary are shown in Figure 4.38. Lag time correlations were conducted
for April to mid-May (time of peak spring flow) during the four years of interest, as well as the
long-term average (1991 - 2020 reference period). Due to data limitations, lags were calculated
with a spatial resolution of days rather than hours. With p values < 0.05, all correlations during
the 1991 - 2020 period were statistically significant. The stations nearest to the Saint John gauge
were most highly correlated (maximum correlation being 0.99 at Oak Point) and—with lag times
of zero days—have the fastest water transport times, indicating that water generally takes less
than a day to travel from the Gagetown gauge to the Saint John gauge. The station least
correlated to the Saint John gauge was Edmundston, which could be due to the influence of the
Madawaska Dam and Hydroelectric Generating Station. The longest water transport times were
those at the Ninemile and Dickey gauges in the upper basin, with average travel times of three
days.
The Ninemile and Dickey gauges had missing data for the first half of April and—due to
the reduced sample size—this could have led to the statistically insignificant correlation
coefficients that were produced. In 2019 and 2021, several of the lower correlation coefficients
generated for high lag times were also statistically insignificant (p-value ≥ 0.05). Of the four years,
2018 was most similar to the long-term average, with the only difference observed at the Dickey
gauge (three days versus two days). The highest and lowest water transport times were in 2008
and 2021, respectively (Figure 4.38), which could be due to water level, as 2008 had the highest

101

water level in both the upper and lower basins whereas 2021 (the non-flood year) had the lowest
(Table 5.1).

102

Figure 4.38 Lag time correlations (0 - 10 days) between eight main stem hydrometric gauges and
water level at the Saint John, NB gauge (near the Saint John River estuary), including the longterm average (1991 – 2020 reference period) and from 1 April to 15 May for the four years of
interest. Gauges are ordered from farthest (Ninemile) to closest (Oak Point) to the Saint John
gauge. Maximum lag values are labelled by the corresponding integer, and red circles identify
correlation coefficients that are not statistically significant (p-value ≤ 0.05).
103

Chapter 5 Discussion and Synthesis
This study aimed to identify the sequence of events—at various spatial and temporal
scales—that led to the 2008, 2018, and 2019 major spring floods in the SJRB, and to compare
these major flood events to the 2021 season, in which no major spring flooding occurred. Results
presented in the previous sections show some commonalities and differences among flood years
(e.g., spring precipitation anomalies in the upper and middle basins, respectively), as well as
commonalities and differences between flood and non-flood years (e.g., total and solid
precipitation anomalies, respectively).
In this section, the focus is synthesising features of the flood years, contrasting these with
the non-flood year and, finally, comparing all these findings against previous studies.
5.1 Flood Years (2008, 2018, 2019)
5.1.1 Chronology
For the three floods, hydrographs of flow or water level—recorded at five hydrometric
stations along the SJR main stem—showed consistencies in timing. Increases began in mid to late
April, and these occurred slightly earlier (approximately a week) in 2019 than in the other two
years (Figure 4.37). Table 5.1 summarises key dates and values for the four years, including flood
duration, flood wave travel times, peak spring flow or level dates, maximum flood stage
exceedance, and number of rainfall events exceeding threshold.
Consistencies in timing also occurred in terms of highest values, with peak values in 2019
occurring approximately a week earlier than the other flood years. Peak flow along the SJR main
stem occurred in the headwaters at Ninemile Bridge in late April (30 April 2008 and 2018; 23 April
104

2019), with peak water level occurring at the mouth at Saint John in late April to early May (4
May 2008; 7 May 2018; 26 April 2019). Water levels recorded at the Edmundston hydrometric
station reached a maximum flood stage exceedance of 3.89 m (1 May 2008), 1.95 m (1 May 2018),
and 2.77 m (19 April 2019). Downstream of Edmundston, water levels recorded at the
Fredericton hydrometric station reached a maximum flood stage exceedance of 1.79 m (1 May
2008), 1.72 m (1 May 2018), and 1.76 m (23 April 2019).
The time evolution of peak flow or water level also showed consistencies as well as
differences (Figure 5.1). This figure was developed following the approach of Blöschl et al. (2013).
With a total peak time lag of 150 h (or 6.25 days), the propagation of the 2018 flood was slower
than the other two flood events. Downstream of the Mactaquac Dam, a relatively sharp increase
in flood propagation—especially evident in 2018—occurred between the Fredericton and
Maugerville gauges during all three floods. Several factors could contribute to this sharp increase
in lag time, including the influence of the Mactaquac Dam, the Grand Lake system (which
provides flood water storage), and the Reversing Falls (which restricts outflow to the Bay of
Fundy).

105

2

1

4

0

5

4

4

1.76

2.77

26 April

23 April

wet-cold

wet-cold

wet-cold

wet-cold

3

4

1

NA

NA

3 April

11 April

dry-warm

dry-warm

dry-warm

wet-warm

NA
NA

NA

2021

106

As stated in the Flood History Database (NBDELG, n.d.-a).
None of the four years was significantly different from the mean precipitation or air temperature (reference period: 1991-2020).

4

1.72

1

1.79

lower basin
(SJR at Fredericton)

1.95

4

3.89

upper basin
(SJR at Edmundston)

7 May

30 April

wet-warm

wet-cold

wet-warm

wet-warm

2

4 May

wet-cold

spring

mouth
(SJR at Saint John)

wet-warm

winter

30 April

dry-cold

wet-warm

spring

winter

headwaters
(SJR at Ninemile Bridge)

upper basin

full basin

total travel time (h)
velocity (km h-1)

upper basin
rainfall events exceeding
threshold
middle basin
(1 April - 15 May)
lower basin

maximum spring flood
stage exceedance (m)

date of peak spring flow
(headwaters) or stage
(mouth)

seasonal precipitation +
temperature anomaly
classifications2

flood wave (h)

duration of flood1

2008
2018
2019
10 days
16 days
17 days
(23 April - 2 May) (27 April - 12 May) (19 April - 6 May)
105
150
76
5.25
3.67
7.25

Table 5.1 Summary of key dates and values for the flood (2008, 2018, 2019) and non-flood (2021) years.

Figure 5.1: Travel times of peak flow or water level for the 2008, 2018, and 2019 spring floods
along the Saint John River main stem, from Ninemile Bridge (plotted as zero) in the headwaters
to Saint John at the estuary. Main stem hydroelectric dams are labelled and identified by
vertical lines.

During the flood years, two or more heavy rainfall events (thresholds defined in Table 4.2)
coincided with—and contributed to—the high peak spring flows and water levels. During the
period from 1 April to 15 May, the 2008 season had two rainfall events that exceeded the heavy
liquid precipitation thresholds in the middle and / or upper basins (Figure 5.2); the 2018 season
had four rainfall events that exceeded the heavy liquid precipitation thresholds in all sub-sections
of the basin (Figure 5.3); and the 2019 season had five rainfall events that exceeded the heavy
liquid precipitation thresholds—four events that exceeded thresholds in all sub-sections of the
basin and one that exceeded the lower basin threshold (Figure 5.4).

107

Figure 5.2 Heavy rainfall events and related storms for 1 April - 15 May 2008. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.

108

Figure 5.3 Heavy rainfall events and related storms for 1 April - 15 May 2018. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.

109

Figure 5.4 Heavy rainfall events and related storms for 1 April - 15 May 2019. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.

110

5.1.2 Seasonal Patterns
During winter, when averaged over the full basin, the flood years all had more
precipitation than normal, but differed with respect to air temperature. As such, 2008 was
classified as being wet-warm, 2018 was wet-warm (slightly warmer than normal), and 2019 was
wet-cold (Figure 4.22; Table 5.1). This pattern holds true for the subsections of the basin as well,
as the flood years remained in or near the wet-cold quadrant (all being wetter than normal, with
2018 being slightly warmer than normal).
Although patterns emerged during winter, spring precipitation and air temperature
anomalies varied among flood years as—when averaged across the full basin—the flood years
were all classified into different categories: 2008 was dry-cold, 2018 was wet-warm, and 2019
was wet-cold (Figure 4.23; Table 5.1). When averaged across the different subsections of the
basin, some disparity among flood years remains, as 2018 and 2019 stayed within the wet-warm
and wet-cold quadrants, respectively. However, 2008 shifted from dry-cold (winter) to wet-cold
(spring) in the upper basin—the same spring classification as 2019.
Positive monthly SWE anomalies were observed during flood years from December to
May in all sections of the basin—except for 2018, which had negative anomalies throughout the
basin in December, and negative anomalies continuing through to March in the lower basin.
During the flood years, cumulative runoff increased (evident by the steeper slopes in
Figure 4.30) in all sections of the basin around mid to late April. This increase in cumulative runoff
coincided with the onset of flooding, as flood stages were surpassed by mid to late April (Figure
4.37).

111

5.1.3 Rain-on-Snow
A critical issue for flooding is the occurrence of rain-on-snow events, as heavy rainfall in
spring can enhance snowmelt and increase runoff, effectively increasing the risk of flooding. Rainon-snow events occurred in all three flood years. Cumulative rainfall and snowmelt aligned
closely from 1 April to 15 May—the period leading up to and including the spring flood events
(Figure 4.25).
The largest rain-on-snow events for the three flood years were examined. For each of
these, the mean surface temperature, mean SWE, precipitation phase, and water volume are
listed in Table 5.2. The largest rain-on-snow event in 2008 occurred 29-30 April when a heavy
rainfall (shown in yellow on Figure 5.2) event brought 49.9 mm of rain to the upper basin and
34.3 mm to the middle basin. The rainfall from this event fell upon an average SWE of 18.9 cm
and 10.1 cm in the upper and middle basins, respectively. In 2018 there were four heavy rain-onsnow events, the heaviest of which occurred on 26 April (shown in yellow on Figure 5.3). This
event exceeded the heavy rainfall thresholds in each section of the basin, producing 28.1 mm,
25.6 mm, and 22.5 mm of rain in the upper, middle, and lower basins, respectively. The average
SWE during this event was 23.0 cm, 14.4 cm, and 2.2 cm, in the upper, middle, and lower basins,
respectively. Similarly, the 2019 season had numerous rain-on-snow events, the largest having
occurred on 19-20 April (shown in green on Figure 5.4). Like 2018, this 2019 event exceeded the
heavy rainfall thresholds in each section of the basin, producing 42.9 mm, 44.7 mm, and 38.8 mm
of rain that fell upon 27.3 cm, 19.3 cm, and 3.1 cm of SWE in the upper, middle, and lower basins,
respectively.

112

Collectively, the presence of heavy rain-on-snow events contributed to the severity of
these major spring flood events, with the effects of these being reflected in the sharp increases
to cumulative snowmelt (Figure 4.25) and runoff (Figure 4.30). Note that the volume of water
was generally higher from rainfall than from snowmelt in these events, although their relative
contributions varied considerably. For example, the water volume contributions of rainfall to
snowmelt was 67 % to 33 % during the 29-30 April 2008 event and 57 % to 43 % during the 26
April 2018 event (Table 5.2).

113

Precipitation
Event

114

Mean Surface
Precipitation
Water Volume
Mean SWE
Sub-Basin Temperature
(cm)
Total (mm) Rainfall (%) Snowfall (%) Total (km3) Rainfall (%) Snowmelt (%)
(°C)
upper
4.0
18.9
52.7
95
5
1.7
67
33
6.0
10.1
34.4
100
0
1.1
54
46
29-30 Apr 2008 middle
lower
7.7
1.1
15.8
100
0
0.3
73
27
upper
5.6
23.0
28.0
100
0
1.1
57
43
26 Apr 2018
middle
7.8
14.4
25.6
100
0
0.9
49
51
lower
9.5
2.2
22.5
100
0
0.4
77
23
upper
3.7
27.3
46.6
92
8
1.6
63
37
19-20 Apr 2019 middle
5.7
19.3
45.1
99
1
1.0
78
22
lower
7.7
3.1
38.8
100
0
0.7
81
19
upper
2.2
5.0
20.1
36
64
0.4
44
56
1 Apr 2021
middle
5.8
2.8
25.7
93
7
0.6
72
28
lower
8.4
0.6
31.8
100
0
0.5
89
11

Table 5.2 Key features of the heaviest rainfall events from 1 April - 15 May for the flood (2008, 2018, 2019) and non-flood (2021) years.

5.2 Non-Flood Year (2021)
5.2.1 Chronology
During the non-flood year, hydrographs of flow or water level—recorded at five
hydrometric stations along the SJR main stem—showed that peaks occurred relatively early
(Table 5.1; Figure 4.37). Water levels recorded at the Edmundston hydrometric station reached
a maximum of 137.34 m on 11 April (1.70 m below flood stage). Downstream of Edmundston,
water levels recorded at the Fredericton hydrometric station reached a maximum of 6.19 m on
2 April (0.30 m below flood stage). Water levels remained below the flood stage during the entire
spring season in all sections of the basin. These early peak water flows and levels could have
arisen because, during the period from 1 April to 15 May, there were four heavy rainfall events—
the largest of which occurred on 1 April. This rainfall event, which coincided with peak flow and
water level, exceeded the heavy liquid precipitation thresholds in the middle and lower basins,
but not the upper basin (Figure 5.5). This event produced an average of 7.2 mm, 23.9 mm, and
31.8 mm across the upper, middle, and lower basins, respectively. At the time of the event, the
average SWE was 5.0 cm (upper basin), 2.8 cm (middle basin), and 0.6 cm (lower basin). This
heavy rainfall event on 1 April could explain why, during the non-flood year, the peak flow and
water level occurred earlier in the lower basin than in the upper and middle basins, as peak flow
occurred in the headwaters at Ninemile Bridge on 11 April and at the estuary in Saint John on 3
April.

115

Figure 5.5 Heavy rainfall events and related storms for 1 April - 15 May 2021. Left: total daily
rainfall for the upper, middle, and lower Saint John River basin (SJRB). The horizontal red dashed
line identifies heavy rainfall event thresholds for the spring—these are specific to the sub-basins,
with rainfall amounts above the red lines being within the largest 10 % of rainfall events for that
region. Bottom right: low pressure system tracks produced from hourly ERA5 data, with dots
identifying 6-hour increments and stars identifying the start of the low pressure systems based
on the code. The tracks are mapped in reference to the SJRB, identified by the black polygon. Low
pressure systems are linked via colour codes to the corresponding rainfall events (identified on
rainfall plots by vertical coloured lines). Top right: central pressure over time for each of the low
pressure systems.

116

5.2.2 Seasonal Patterns
During the winter, when averaged over the full basin, the non-flood year was wetter and
warmer than normal and was therefore classified as being wet-warm. This pattern is also
observed when viewing the upper, middle, and lower basins separately, with the non-flood year
being wet-warm in the middle and lower basins, but dry-warm (slightly drier than normal) in the
upper basin. Thus, throughout the basin, the non-flood year experienced warmer than normal
temperatures and greater than normal total precipitation⁠—except for the upper basin, which
was slightly drier than normal (Figure 4.22; Table 5.1).
During the spring, the non-flood year was classified as dry-warm when precipitation and
air temperature were averaged over the full, upper, middle, and lower basins (Figure 4.23; Table
5.1).
Regarding monthly anomalies, the non-flood year had positive temperature anomalies
during all winter (December, January, February) and spring (March, April, May) months in all
sections of the basin (Figure 4.16), and negative SWE anomalies in all months and sections except
October in the upper basin, which was slightly positive (Figure 4.20).
The majority of snowmelt occurred in early April (Figures 4.24). Although SWE was low,
this snowmelt coincided with the 1 April rainfall event (which exceeded heavy liquid precipitation
thresholds in the middle and lower basins) and peak spring water level in the lower basin.
5.2.3 Rain-on-Snow
Rain-on-snow events were not a major feature of the 2021 spring season. During this nonflood year, snowmelt largely preceded the rainfall season (Figure 4.25). One exception was the
117

rain-on-snow event that occurred on 1 April when a rainfall event (shown in yellow on Figure 5.5)
brought 7 mm, 24 mm, and 32 mm of rain to the upper, middle, and lower basins, respectively.
However, the average SWE during the non-flood year was relatively low (approximately 90%
below average of normal April SWE), so the rainfall from this event fell upon only 5 cm, 3 cm, and
0.6 cm of SWE in the upper, middle, and lower basins, respectively (Table 5.2). Consequently, in
comparison to the flood years, this event led to the smallest contribution of snowmelt. This rainon-snow event resulted in the sharp increase in cumulative snowmelt (Figure 4.25) in early April,
and during the 1 April to 15 May period the rate of cumulative runoff remained fairly constant
(Figure 4.30).
5.3 Comparisons between Flood and Non-Flood Years
5.3.1 Commonalities
Several features occurred within the non-flood year and one or more of the flood years
(Table 5.3). When averaged over the full basin, all four years had positive total precipitation
anomalies in winter. The non-flood year and two flood years (2018 and 2019) had positive heavy
liquid precipitation anomalies in spring; and the non-flood year and one flood year (2018) had
positive temperature anomalies in spring.
When averaged over the upper basin, the non-flood year was no longer characterised by
heavy liquid precipitation anomalies in spring. However, there are still commonalities found with
the other two parameters: the non-flood year and two flood years (2008 and 2018) had positive
total precipitation anomalies in winter; and the non-flood year and one flood year (2018) had
positive temperature anomalies in spring.
118

5.3.2 Differences
When comparing features of the flood and non-flood years, differences outnumber
commonalities. As shown in Table 5.3, five and six contributing factors were present only in flood
years when averaged over the full and upper basins, respectively. Other differences include ice
jams and high tides, which were listed as causal factors in the Flood History Database for one or
more of the flood years (NBDELG, n.d.-a).
When averaged over the full basin, there were four features that were present during all
three flood years but absent during the non-flood year: positive heavy solid precipitation
anomalies in spring; positive monthly SWE anomalies; steep increases in cumulative runoff in
April; and rain-on-snow events.
When averaged over the upper basin, five features were present during all three flood
years but absent during the non-flood year. These were positive total precipitation anomalies in
winter; positive total precipitation anomalies in spring; positive monthly SWE anomalies; steep
increases in cumulative runoff in April; and rain-on-snow events.

119

causal factor
causal factor
positive anomalies in winter

ice jams

high tides 2

total precipitation

general 1

2

1

steep increases in April
occurrence

cumulative runoff

rain-on-snow

✔
✔

✔

✔

✔
✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

✔

2021

Listed as a causal factor of flooding in the Flood History Database (NBDELG, n.d.-a).
Tidal influences in the SJR extends 130 km upstream⁠—10 km downstream of Mactaquac Dam (Andrews et al., 2017).

positive anomalies in spring

temperature

positive monthly anomalies (Jan-May)

✔

✔

heavy precipitation positive solid precipitation anomalies in winter
events
positive liquid precipitation anomalies in spring

✔

✔

✔

positive anomalies in spring

averaged over
the upper basin SWE

✔

✔

positive anomalies in winter

total precipitation

✔

✔

✔

✔

✔
✔

✔

✔

✔
✔

✔

✔

✔

occurrence

rain-on-snow

✔

✔
✔

2019

2018

✔

steep increases in April

cumulative runoff

✔

✔

✔

2008

✔

positive anomalies in spring

temperature

positive anomalies in spring
heavy precipitation positive solid precipitation anomalies in winter
positive liquid precipitation anomalies in spring
averaged over events
the full basin SWE
positive monthly anomalies (Jan-May)

Contributing Factors

Parameter

Area

Table 5.3 Contributing factors during the flood (2008, 2018, 2019) and non-flood (2021) years.

120

5.4 Comparison with Previous Studies
Flooding is studied across Canada and in many areas of the world, and the cause is often
complex, involving a combination of flood generating mechanisms (Tarasova et al., 2019). Even
for cold climates, numerous pathways lead to flooding⁠. These pathways include different types /
classes of floods, such as meteorological (e.g., heavy rain, extratropical storms); hydrological
(e.g., rain-on-snow, ice jam); and geomorphic (e.g., avalanche related, landslide) (Burrell, 2011;
Buttle et al., 2016).
Across Canada, studies have demonstrated the importance of snowmelt and extreme
precipitation as flood generating mechanisms. One such example of extreme precipitationinduced flooding are the November 2021 floods that occurred in southwestern British Columbia.
The result of back-to-back atmospheric rivers, these floods devastated the region and were “the
costliest natural disaster in the province's history” (Gillett et al., 2022, p. 1). An investigation of
the 2013 Bow River basin flood in Alberta conducted by Teufel et al. (2017) found that the flood
was caused by a combination of meteorological and hydrological factors, with the driving force
having been heavy liquid precipitation. The heavy precipitation fell upon snow in the upper basin,
contributing additional snowmelt-derived runoff. Despite being located in western Canada,
approximately 3400 km away, similarities exist between the 2013 Bow River flood and the three
SJR floods in the present study. Like the Bow River, two of the SJR floods (2018 and 2019) had
heavy liquid precipitation events in the upper basin, with rain-on-snow and increased runoff in
all flood years (Table 5.3).

121

On the East Coast, Singh et al. (2021) found that flooding was primarily driven by
snowmelt. Studies conducted on floods in QC (e.g., Milbrandt and Yau, 2001; Saint-Laurent et al.,
2009) and ME (Budd, 1988) found major spring floods to be caused by heavy rain, rain-on-snow,
and snowmelt-induced runoff. The “April Fools flood of 1987” in ME was especially devastating,
with the Kennebec River basin most severely impacted. Other rivers in ME that also experienced
major flooding include the Penobscot, Saco, and Androscoggin.
According to Burrell (2011), inland flooding in NB is “primarily the result of rain or rainon-snow events, ice jamming, or a combination of these factors” (p. 5). Offering support to this
claim are several studies conducted on the SJR that have found snowmelt and rain-on-snow
events to be generating mechanisms of major spring floods (e.g., Buttle et al., 2016; Inland
Waters Directorate, 1974; Newton & Burrell, 2016). The importance of these flood generating
mechanisms has also been communicated by local professionals (J. Doiron, personal
communication, 10 June 2021). With rain-on-snow events and snowmelt (exhibited by high SWE
and cumulative runoff rates) present in the three flood years, the present findings further
illustrate the importance of these features in generating major spring flooding in the SJRB.
Two previous studies that specifically focused on the 2008 flood were Lombard (2010)
and Newton and Burrell (2016). Aligning with the present study, both previous studies listed the
generating mechanisms as being an above average snowpack in April, rapid snowmelt, and
rainfall. These same factors were identified in the present study, but the previous studies had
not considered actual storm tracks and precipitation events, nor did they consider sub-basin
averages—the previous studies used meteorological data from weather stations (point data)
rather than gridded, reanalysis data.
122

Although studies have identified snowmelt and rain-on-snow as flood generating
mechanisms, many of the studies conducted on the SJR focused primarily on ice jam-related
flooding and the conditions that govern ice jam processes (e.g., Beltaos, 1999; Beltaos et al.,
2003; Beltaos & Prowse, 2001). The present findings are similar to these previous SJR studies, in
that ice jams were confirmed by NBDELG (n.d.-a) to be a causal factor in two of the major spring
flood events. In addition to the SJR, the Exploits River in Newfoundland is an example of another
Atlantic Canada river in which “spectacular ice jams may now occur in midwinter, instantly
flooding and freezing riverside towns” (Cunjak & Newbury, 2005, p. 943).
In researching the climatic effects on the changing ice-breakup regime of the SJR, Beltaos
(1999) found an association between an increase in April rainfall and increasing freshet. This
association was seen in the present study in the 2018 and 2019 major spring floods, with April
2019 rainfall anomalies reaching 60 % when averaged over the full basin and 100 % in the lower
basin (Figure 4.18). However, negative rainfall anomalies occurred in the remaining flood year
(2008), whereas the non-flood year had positive anomalies (as much as 20 % in the lower basin).

123

Chapter 6 Concluding Remarks
6.1 Summary
Using multiple perspectives, this study addressed the following two objectives: (1)
identify the sequence of events—at various spatial and temporal scales—that led to the 2008,
2018, and 2019 major spring floods in the SJRB, and (2) compare these major flood events to the
2021 season, in which no major spring flooding occurred. Although the 2008 flood has been the
exclusive focus of previous studies (e.g., Lombard, 2010; Newton & Burrell, 2016), the current
study is, to the best of my knowledge, the first to analyse the conditions that led to the 2018 and
2019 major spring flood events, and the first to compare conditions between spring flood and
non-flood years on the SJR. The results obtained in this research confirm what was known as
causal factors for the 2008, 2018, and 2019 floods.
As this research relied heavily upon reanalysis datasets, evaluations of key variables—air
temperature, precipitation, snow depth, and SWE—were carried out, and these compared
reasonably well to the in-situ weather station and snow survey data. Given this, these tools
represent a suitable retrospective analysis dataset to examine these events.
A comparison of the conditions that led to the three major spring floods revealed
commonalities and differences (Table 5.3). Among the flood years, there was consistency in the
timing of peak flow and water level, with increases that began in mid to late April. Other
commonalities include higher than average winter and spring precipitation in the upper basin;
positive solid precipitation anomalies in winter when averaged over the full SJRB; positive SWE
anomalies (January to May) when averaged over the full and upper basins; steep increases
cumulative runoff in April when averaged over the full and upper basins; and the occurrence of
124

rain-on-snow events. Some features did not occur in all flood years, including ice jams (two years)
and positive spring total precipitation anomalies when averaged over the full SJRB (two years).
A comparison of the conditions between flood and non-flood years also revealed
commonalities and differences (Figure 5.3). Differences in geopotential height anomalies
between the three flood years and the one non-flood year occurred in December (mostly
negative during flood years; positive during non-flood year) and January (mostly positive during
flood years; positive to the north and negative to the southeast of SJRB during non-flood year).
Unlike the flood years, peak flow and water levels occurred in early April during the non-flood
year. Other notable differences include SWE anomalies, which were largely positive during the
flood years and negative during the non-flood year, as well as the influence of rain-on-snow
events which, due to the low SWE values, were inconsequential during the non-flood year.
There are also common features between the non-flood year and one or more of the
flood years. These commonalities include those pertaining to the full basin (positive winter total
precipitation anomalies; positive heavy liquid precipitation anomalies in spring; and positive
temperature anomalies in spring) and the upper basin (positive temperature anomalies in winter
and spring). Commonalities also exist regarding storm tracks, with most tracks having a SW to NE
direction of movement that was independent of month and whether it was a flood or non-flood
year. Although the directionality was similar, there are clear distinctions between the flood and
non-flood years with respect to the number of tracks, as the non-flood year had the fewest total
storm tracks and zero tracks that passed north of the basin.
These findings are largely consistent with those of previous studies. An above average
April snowpack and rapid snowmelt normally enhanced by major spring rain events are common
125

features. However, this study more thoroughly quantified and interpreted the spatial and
temporal characteristics of these features, both within the same flood year and between flood
years. Although some of the same features occurred in all flooding years, there was substantial
variability in their magnitude, location, and timing.
In summary, this research illustrates that flooding in the SJRB is the result of a complex
combination of causative factors that influence the timing and magnitude of major flood events,
such as those that occurred in 2008, 2018, and 2019. The concurrence, or lack thereof, of key
meteorological conditions is a critical aspect of this flooding. In particular, rain-on-snow events
were a prominent feature of the three flood years but not in 2021, not because there were no
rainstorms but because there was low SWE when they occurred due to an early snowmelt.
6.2 Future work
6.2.1 Additional Analyses
Although comparisons were conducted to confirm the similarity between reanalysis and
in-situ data, the results generated for the hydroclimatic and hydrologic perspectives (Sections 4.1
and 4.2, respectively) relied exclusively on reanalysis data. Future analyses could incorporate insitu data, especially for attributes of the hydrologic perspective, for which weather station data
are available. Remote sensing and model simulations could also provide additional resources to
examine floods in the SJRB, with models being especially useful for looking at impacts of potential
future climate change on flooding (e.g., Deb & Laroche, 2022).
Although many flood generating mechanisms were explored and identified, there were
others that fell outside the scope of this project. From a hydroclimatic perspective, performing a

126

synoptic climatology would constitute another method of linking climatic and weather systems
to flood events; this could follow the work of Yarnal (1993). An additional approach to flood
classification from the hydrological perspective could include the use of clustering to group flood
events and derive flood types through inductive analysis (e.g., Baronetti et al., 2019; Keller et al.,
2018).
During rain-on-snow events in the upper SJRB, some events also produced snowfall (Table
5.2). The occurrence of snow, rainfall, and mixed precipitation would imply that the surface air
temperature was near 0 °C somewhere within this region. Future work should identify the factors
governing the local and regional location of this isotherm. Such an investigation could utilise the
detailed precipitation measurements and analyses of the SAJESS field campaign and expand
these to somewhat larger scales.
In this study, only one non-flood year was included (2021, the year during which the
SAJESS field campaign was conducted). Future studies should include additional flood and nonflood years to reveal patterns that were not evident in the present study and, conversely, could
yield results that conflict with the present findings. For example, perhaps other non-flood years
were characterised by a lack of spring rain events or gradual snowmelt, not by a lack of SWE such
as occurred in 2021. Increasing the number of flood and non-flood years would also enable more
statistical analyses of anomalies to be conducted, providing a more comprehensive perspective
in the patterns that were discerned in this study.
6.2.2 Flood Preparedness and Mitigation
The impacts of flooding on society depend somewhat on how prepared communities are.
For example, the 2018 flood exceeded the other two floods in the number of properties affected,
127

as well as the estimated cost of damages. Flood damages estimated by NBDELG (n.d.-a) exceeded
$23 million in 2008 and $75 million in 2018; no comparable figure is known for 2019, as it has not
yet been listed in the NBDELG Flood History Database, but it would be considerably lower than
in 2018. After the 2019 flood, the NB government received approximately one third of the
damage reports and financial assistance applications when compared to 2018. According to
NBDELG (n.d.-a), this decrease was attributed to “more advanced warning, preparations and
lessons learned from 2018 helped lessen flood impacts and costs in 2019.” Perhaps the findings
from this study could better guide future preparations.
Flood events in the SJRB will continue to occur and, due to climate change, could become
more prevalent than ever (Hare et al., 1997). As such, in addition to becoming better prepared
for flood events, flood mitigation needs to be addressed (Duhamel et al., 2022). When exploring
potential flood mitigation in the SJRB, forestry—a major industry in NB—is one of the factors that
should be considered. This would include assessing the extent to which forest mortality and
harvesting may be contributing to or mitigating the peak flow at designated locations of high
flood risk. The resulting insight could provide suggestions as to how to reduce flooding.

128

References
Ahrens, C. D., Jackson, P. L., & Jackson, C. E. J. (2016). Meteorology today: An introduction to
weather, climate, and the environment (2nd ed.). Nelson Education Ltd.
Andrews, S., Linnansaari, T., Curry, R., & Dadswell, M. (2017). The misunderstood Striped Bass
of the Saint John River, New Brunswick: Past, present and future. North American
Journal of Fisheries Management, 37, 235–254.
https://doi.org/10.1080/02755947.2016.1238424
Atlantic Coastal Action Program Saint John (2020). Understanding climate change in Saint John,
2020 report. https://saintjohn.ca/en/city-hall/corporate-plans-and-projects/climatechange-action-plans
Baronetti, A., Fratianni, S., Acquaotta, F., & Fortin, G. (2019). A quality control approach to
better characterize the spatial distribution of snow depth over New Brunswick,
Canada. International Journal of Climatology, 39(14), 5470–5485.
https://doi.org/10.1002/joc.6166
Beltaos, S. (1999). Climatic effects on the changing ice-breakup regime of the Saint John River.
Proceedings of the 10th Workshop on the Hydraulics of Ice Covered Rivers. CRIPE.
http://www.cripe.ca/publications/proceedings/10
Beltaos, S., Ismail, S., & Burrell, B. C. (2003). Midwinter breakup and jamming on the upper
Saint John River: A case study. Canadian Journal of Civil Engineering, 30(1), 77–88.
https://doi.org/10.1139/l02-062
Beltaos, S., & Prowse, T. D. (2001). Climate impacts on extreme ice-jam events in Canadian
rivers. Hydrological Sciences Journal, 46(1), 157–181.
https://doi.org/10.1080/02626660109492807
Blöschl, G., Nester, T., Komma, J., Parajka, J., & Perdigão, R. A. P. (2013). The June 2013 flood in
the Upper Danube Basin, and comparisons with the 2002, 1954 and 1899 floods.
Hydrology and Earth System Sciences, 17(12), 5197–5212. https://doi.org/10.5194/hess17-5197-2013
Brimelow, J., Szeto, K., Bonsal, B., Hanesiak, J., Kochtubajda, B., Evans, F., & Stewart, R. (2015).
Hydroclimatic aspects of the 2011 Assiniboine River Basin flood. Journal of
Hydrometeorology, 16(3), 1250–1272. https://doi.org/10.1175/JHM-D-14-0033.1
Budd, B. A. (1988). The 1987 April Fools flood in Maine: A hydro-meteorological account. Natl.
Wea. Dig., 13(4), 6-15. https://nwas.org/publications/digest/
129

Burrell, B. (2011). Inland flooding in Atlantic Canada. Atlantic Climate Adaptation Solutions
Association. https://atlanticadaptation.ca/en/islandora/object/acasa%3A484
Buttle, J.M., Allen, D. M., Caissie, D., Davison, B., Hayashi, M., Peters, D. L., Pomeroy, J. W.,
Simonovic, S., St-Hilaire, A., & Whitfield, P. H. (2016). Flood processes in Canada:
Regional and special aspects. Canadian Water Resources Journal / Revue Canadienne
Des Ressources Hydriques, 41(1–2), 7–30.
https://doi.org/10.1080/07011784.2015.1131629
CBC News. (2018, May 7). The one chart you need to see about the New Brunswick flood. CBC.
https://www.cbc.ca/news/canada/new-brunswick/new-brunswick-flood-2018-2008water-levels-1.4652494
Chartrand, J., & Pausata, F. S. R. (2020). Impacts of the North Atlantic Oscillation on winter
precipitations and storm track variability in southeast Canada and the northeast United
States. Weather and Climate Dynamics, 1(2), 731–744. https://doi.org/10.5194/wcd-1731-2020
Chen, T.-C., Di Luca, A., Winger, K., Laprise, R., & Thériault, J. M. (2022). Seasonality of
continental extratropical-cyclone wind speeds over northeastern North America.
Geophysical Research Letters, 49(15), e2022GL098776.
https://doi.org/10.1029/2022GL098776
Clair, T., Beltaos, S. Brimley, & Diamond, A. (1997). Climate change and climate variability in
Atlantic Canada. In J. Abraham, T. Canavan & R. Shaw (Eds.), The Canada country study:
Climate impacts and adaptation. Environment Canada.
https://www.researchgate.net/publication/308518282_Climate_Change_and_Climate_
Variability_in_Atlantic_Canada
CoCoRaHS. (n.d.). CoCoRaHS data reports. https://cocorahs.org/ViewData/
Cunjak, R. A., & Newbury, R. W. (2005). Atlantic coast rivers of Canada. In A.C. Benke & C.E.
Cushing (Eds.), Rivers of North America (pp. 938–980). Elsevier.
https://doi.org/10.1016/B978-012088253-3/50024-9
Deb, S., & Laroche, A. M. (2022). Floodplain mapping of Nashwaak River, Canada: An
assessment based on high-resolution hydraulic simulation. Journal of Shipping and
Ocean Engineering, 12, 1–11. https://doi.org/10.17265/2159-5879/2022.01.001
Derksen, C., Burgess, D., Duguay, C., Howell, S., Mudryk, L., Smith, S., Thackeray, C., &
Kirchmeier-Young, M. (2019). Changes in snow, ice, and permafrost across Canada. In E.
Bush and D.S. Lemmen (Eds.), Canada’s Changing Climate Report (pp. 194–260).
Government of Canada.
130

Déry, S. J., & Wood, E. F. (2005). Observed twentieth century land surface air temperature and
precipitation covariability. Geophysical Research Letters, 32(L21414).
https://doi.org/10.1029/2005GL024234
Duhamel, F., Germain, D., & Fortin, G. (2022). Flood risk quantification and mapping: An
integrated vulnerability-based approach at the local scale in New Brunswick, Canada.
International Journal of Disaster Risk Reduction, 83, 103403.
https://doi.org/10.1016/j.ijdrr.2022.103403
Environment and Climate Change Canada. (2017, August 8). Top ten weather stories for 2008:
story four: Saint John River Floods from top to bottom. https://ec.gc.ca/meteoweather/default.asp?lang=En&n=7D6FDB7C-1
Environment and Climate Change Canada. (2019, January 10). Canada’s top 10 weather stories
of 2018: 7. Flash flooding of the Saint John River.
https://www.canada.ca/en/environment-climate-change/services/top-ten-weatherstories/2018.html#toc6
Environment and Climate Change Canada. (2020, January 8). Canada’s top 10 weather stories of
2019: 9. Saint John River floods again. https://www.canada.ca/en/environment-climatechange/services/top-ten-weather-stories/2019.html#toc10
Environment and Climate Change Canada. (2021, November 16). 1981-2010 Climate normals &
averages. https://climate.weather.gc.ca/climate_normals/index_e.html
El-Jabi, N., Turkkan, N., & Caissie, D. (2013). Regional climate index for floods and droughts
using Canadian Climate Model (CGCM3.1). American Journal of Climate Change, 02(02),
106–115. https://doi.org/10.4236/ajcc.2013.22011
Elliott, J. G., Jarrett, R. D., & Ebling, J. L. (1982). Annual snowmelt and rainfall peak-flow data on
selected foothills region streams, South Platte River, Arkansas River, and Colorado River
Basins, Colorado (Report No. 82-426). U.S. Geological Survey.
https://doi.org/10.3133/ofr82426
Fortin, G., & Dubreuil, V. (2020). A geostatistical approach to create a new climate types map at
regional scale: Case study of New Brunswick, Canada. Theoretical and Applied
Climatology, 139(1–2), 323–334. https://doi.org/10.1007/s00704-019-02961-2
Gillett, N. P., Cannon, A. J., Malinina, E., Schnorbus, M., Anslow, F., Sun, Q., Kirchmeier-Young,
M., Zwiers, F., Seiler, C., Zhang, X., Flato, G., Wan, H., Li, G., & Castellan, A. (2022).
Human influence on the 2021 British Columbia floods. Weather and Climate Extremes,
36(100441), 1–13. https://doi.org/10.1016/j.wace.2022.100441

131

Government of Canada. (2019, October 30). Floods. Get Prepared.
https://www.getprepared.gc.ca/cnt/hzd/flds-en.aspx
Government of Canada. (2021a, July 9). Historical data.
https://climate.weather.gc.ca/historical_data/search_historic_data_e.html
Government of Canada. (2021b, July 27). Terrestrial ecoregions of Canada.
https://open.canada.ca/data/en/dataset/ade80d26-61f5-439e-8966-73b352811fe6
Government of Canada. (2021c, July 28). Canadian historical snow survey data.
https://open.canada.ca/data/en/dataset/f57afb37-90c7-4ff0-a757-ba792a4f40a1
Gupta, H. V., Kling, H., Yilmaz, K. K., & Martinez, G. F. (2009). Decomposition of the mean
squared error and NSE performance criteria: Implications for improving hydrological
modelling. Journal of Hydrology, 377(1), 80–91.
https://doi.org/10.1016/j.jhydrol.2009.08.003
Hare, F.K., Dickison, R.B.B., & Ismail, S. (1997, September). Variations of climate and streamflow
over the Saint John Basin since 1872 [Conference], The 9th Workshop on River Ice,
Fredericton, NB, Canada. https://www.osti.gov/etdeweb/biblio/607081
Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J.,
Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D., Thépaut, J-N.
(2019). ERA5 monthly averaged data on pressure levels from 1979 to present [Dataset].
Copernicus Climate Change Service (C3S) Climate Data Store (CDS). (Accessed on 23
September 2021), 10.24381/cds.6860a573
Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J.,
Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X.,
Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., … Thépaut, J. (2020). The
ERA5 global reanalysis. Quarterly Journal of the Royal Meteorological Society, 146(730),
1999–2049. https://doi.org/10.1002/qj.3803
Hirschboeck, K. K. (1987). Hydroclimatically-defined mixed distributions in partial duration flood
series. In V. P. Singh (Ed.), Hydrologic Frequency Modeling (pp. 199–212). Springer
Netherlands. https://doi.org/10.1007/978-94-009-3953-0_13
Hirschboeck, K. K. (1988). Flood hydroclimatology. In V. R. Baker, R. C. Kochel, & P. C. Patton
(Eds.), Flood geomorphology (pp. 27–49). Wiley.

132

House, P. K., & Hirschboeck, K. K. (1997). Hydroclimatological and paleohydrological context of
extreme winter flooding in Arizona, 1993. In R. A. Larson & J. E. Slossn (Eds.), Storminduced geological hazards: Case histories from the 1992–1993 winter storm in southern
California and Arizona, (pp. 1–24). Geological Society of America.
https://doi.org/10.1130/REG11-p1
Inland Waters Directorate. (1974). New Brunswick flood April–May 1973 (Technical Bulletin No.
81). Environment Canada.
https://publications.gc.ca/site/eng/9.861851/publication.html
Intergovernmental Panel on Climate Change. (2019). Summary for policymakers. In: Climate
change and land: An IPCC special report on climate change, desertification, land
degradation, sustainable land management, food security, and greenhouse gas fluxes in
terrestrial ecosystems [P.R. Shukla, J. Skea, E. Calvo Buendia, V. Masson-Delmotte, H.- O.
Pörtner, D. C. Roberts, P. Zhai, R. Slade, S. Connors, R. van Diemen, M. Ferrat, E.
Haughey, S. Luz, S. Neogi, M. Pathak, J. Petzold, J. Portugal Pereira, P. Vyas, E. Huntley,
K. Kissick, M. Belkacemi, J. Malley, (eds.)]. https://doi.org/10.1017/9781009157988.001
International Joint Commission. (1975, January 1). Water quality in the Saint John River basin.
https://ijc.org/en/water-quality-saint-john-river-basin
IRI. (n.d.). IRI Climate and Society map room. Retrieved September 30, 2022, from
https://iridl.ldeo.columbia.edu/maproom/index.html
Jemseg Grand Lake Watershed Association. (2021, February 1). Flooding in New Brunswick:
Cause and effect [Video]. YouTube.
Keller, L., Rössler, O., Martius, O., & Weingartner, R. (2018). Delineation of flood generating
processes and their hydrological response. Hydrological Processes, 32(2), 228–240.
https://doi.org/10.1002/hyp.11407
Kidd, S.D., Curry, R.A., & Munkittrick, K.R. (Eds.). (2011). The Saint John River: A state of the
environment report. Canadian Rivers Institute, UNB.
Lombard, P. J. (2010). Flood of April and May 2008 in Northern Maine: U.S. Geological Survey
scientific investigations report 2010–5003. U.S. Geological Survey.
http://pubs.usgs.gov/sir/2010/5003/
Mai, J., Shen, H., Tolson, B. A., Gaborit, É., Arsenault, R., Craig, J. R., Fortin, V., Fry, L. M., Gauch,
M., Klotz, D., Kratzert, F., O’Brien, N., Princz, D. G., Rasiya Koya, S., Roy, T., Seglenieks, F.,
Shrestha, N. K., Temgoua, A. G. T., Vionnet, V., & Waddell, J. W. (2022). The Great Lakes
Runoff Intercomparison Project Phase 4: The Great Lakes (GRIP-GL). Hydrology and
Earth System Sciences, 26(13), 3537–3572. https://doi.org/10.5194/hess-26-3537-2022
133

Maine GeoLibrary. (2017). Maine Biophysical Regions GeoLibrary Poly [Dataset]. Maine GIS Data
Catalog. https://www.maine.gov/geolib/catalog.html
Mekis, E., & Hogg, W. D. (1999). Rehabilitation and analysis of Canadian daily precipitation time
series. Atmosphere-Ocean, 37(1), 53–85.
https://doi.org/10.1080/07055900.1999.9649621
Milbrandt, J. A., & Yau, M. K. (2001). A mesoscale modeling study of the 1996 Saguenay Flood.
Monthly Weather Review, 129(6), 1419–1440. https://doi.org/10.1175/15200493(2001)129%3C1419:AMMSOT%3E2.0.CO;2
Muñoz-Sabater, J. (2019). ERA5-Land hourly data averaged data from 1981 to present
[Dataset]. Copernicus Climate Change Service (C3S) Climate Data Store (CDS). (Accessed
on 23 September 2021), 10.24381/cds.68d2bb3
Muñoz-Sabater, J., Dutra, E., Agustí-Panareda, A., Albergel, C., Arduini, G., Balsamo, G.,
Boussetta, S., Choulga, M., Harrigan, S., Hersbach, H., Martens, B., Miralles, D. G., Piles,
M., Rodríguez-Fernández, N. J., Zsoter, E., Buontempo, C., & Thépaut, J.-N. (2021). ERA5Land: A state-of-the-art global reanalysis dataset for land applications. Earth System
Science Data, 13(9), 4349–4383. https://doi.org/10.5194/essd-13-4349-2021
National Weather Service. (n.d.). National Weather Service. https://www.weather.gov/
New Brunswick Department of Environment and Local Government. (n.d.-a). Flood History
Database. Government of New Brunswick.
https://www.elgegl.gnb.ca/0001/en/Flood/Search?LocationName=Saint+John+River+%
28watershed%29
New Brunswick Department of Environment and Local Government. (n.d.-b). Climate change in
New Brunswick. Government of New Brunswick.
https://www2.gnb.ca/content/gnb/en/departments/elg/environment/content/climate_
change/content/climate_change_affectingnb.html
New Brunswick Department of Environment and Local Government. (n.d.-c). Flooding in New
Brunswick. Government of New Brunswick.
https://www2.gnb.ca/content/gnb/en/departments/elg/environment/content/flood.ht
ml
New Brunswick Department of Environment and Local Government. (2014). New Brunswick’s
flood risk reduction strategy. Government of New Brunswick.
https://www2.gnb.ca/content/dam/gnb/Departments/env/pdf/FloodingInondations/NBFloodRiskReductionStrategy.pdf

134

New Brunswick Department of Natural Resources. (2007). Enduring features: Geology and
topography, climate and soil. In Vincent F. Zelazny (Ed.), Our landscape heritage: The
story of ecological land classification in New Brunswick (pp. 21-38). Government of New
Brunswick.
Newton, B., & Burrell, B. C. (2016). The April–May 2008 flood event in the Saint John River
Basin: Causes, assessment and damages. Canadian Water Resources Journal / Revue
Canadienne Des Ressources Hydriques, 41(1–2), 118–128.
https://doi.org/10.1080/07011784.2015.1009950
Roy, P. & Huard, D. (2016, January). Future climate scenarios province of New-Brunswick
(Project No. 50001). Ouranos.
Royé, D. (2020, March 29). Visualize climate anomalies [R Code].
https://dominicroye.github.io/en/2020/visualize-climate-anomalies/
Saint-Laurent, D., Mesfioui, M., & Evin, G. (2009). Hydroclimatic variability and relation with
flood events (Southern Québec, Canada). Water Resources, 36(1), 43–56.
https://doi.org/10.1134/S0097807809010047
Saint John River Experiment on Cold Season Storms. (n.d.). Saint John River Experiment on Cold
Season Storms (SAJESS). https://gwf-sajess.weebly.com/
Sandink, D., Kovacs, P., Oulahen, G., & McGillivray, G. (2010). Making floods insurable for
Canadian homeowners: A discussion paper. Institute for Catastrophic Loss Reduction &
Swiss Reinsurance Company. https://www.iclr.org/wp-content/uploads/PDFS/makingflood-insurable-for-canadian-homeowners.pdf
Shelton, M. L. (2009). Hydroclimatology: Perspectives and applications. Cambridge University
Press.
Singh, J., Ghosh, S., Simonovic, S. P., & Karmakar, S. (2021). Identification of flood seasonality
and drivers across Canada. Hydrological Processes, 35(10), 1–11.
https://doi.org/10.1002/hyp.14398
Stewart, R. E., Bachand, D., Dunkley, R. R., Giles, A. C., Lawson, B., Legal, L., Miller, S. T.,
Murphy, B. P., Parker, M. N., Paruk, B. J., & Yau, M. K. (1995). Winter storms over
Canada. Atmosphere-Ocean, 33(2), 223–247.
https://doi.org/10.1080/07055900.1995.9649533
Swansburg, E., El-Jabi, N., & Caissie, D. (2004). Climate change in New Brunswick (Canada):
statistical downscaling of local temperature, precipitation, and river discharge (Report
No. 2544). Canadian Technical Report of Fisheries and Aquatic Sciences.
135

Tarasova, L., Merz, R., Kiss, A., Basso, S., Blöschl, G., Merz, B., Viglione, A., Plötner, S., Guse, B.,
Schumann, A., Fischer, S., Ahrens, B., Anwar, F., Bárdossy, A., Bühler, P., Haberlandt, U.,
Kreibich, H., Krug, A., Lun, D.,... Wietzke, L. (2019). Causative classification of river flood
events. WIREs Water, 6(4), 1-26. https://doi.org/10.1002/wat2.1353
Tarek, M., Brissette, F. P., & Arsenault, R. (2020). Evaluation of the ERA5 reanalysis as a
potential reference dataset for hydrological modelling over North America. Hydrology
and Earth System Sciences, 24(5), 2527–2544. https://doi.org/10.5194/hess-24-25272020
Teufel, B., Diro, G. T., Whan, K., Milrad, S. M., Jeong, D. I., Ganji, A., Huziy, O., Winger, K.,
Gyakum, J. R., de Elia, R., Zwiers, F. W., & Sushama, L. (2017). Investigation of the 2013
Alberta flood from weather and climate perspectives. Climate Dynamics, 48(9–10),
2881–2899. https://doi.org/10.1007/s00382-016-3239-8
United States Geological Survey. (2013). Shuttle Radar Topography Mission (SRTM) 1 ArcSecond Global [dataset]. https://doi.org/10.5066/F7PR7TFT
United States Geological Survey. (2021, September 23). USGS surface-water daily data for the
nation. https://nwis.waterdata.usgs.gov/nwis/dv
Water Survey of Canada. (2021, February 12). Historical hydrometric data. Government of
Canada. https://wateroffice.ec.gc.ca/mainmenu/historical_data_index_e.html
Whitfield, P. H. (2012). Floods in future climates: A review. Journal of Flood Risk Management,
5(4), 336–365. https://doi.org/10.1111/j.1753-318X.2012.01150.x
Yarnal, B. (1993). Synoptic climatology in environmental analysis: A primer. London, England:
Belhaven Press.
Zhang, X., Flato, G., Kirchmeier-Young, M., Vincent, L., Wan, H., Wang, X., Rong, R., Fyfe, J., Li,
G., & Kharin, V.V. (2019). Changes in temperature and precipitation across Canada. In E.
Bush & D.S. Lemmen (Eds.), Canada’s Changing Climate Report (pp. 112-193).
Government of Canada.
Zhang, X., Hogg, W. D., & Mekis, É. (2001). Spatial and temporal characteristics of heavy
precipitation events over Canada. Journal of Climate, 14, 1923-1936.
https://doi.org/10.1175/1520-0442(2001)014<1923:SATCOH>2.0.CO;2

136

Appendices
A.1 Monthly Geopotential Height Anomaly

137

2008

2018

2019

2021

Figure A.1 December geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) of the four water years of interest, mapped
on a 2.5° grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 30 geopotential
metres (gpm). Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown in shades
of blue and are also expressed in units of gpm. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 2020 reference period (IRI, n.d.).
138

925
hPa

500
hPa

250
hPa

Dec

2008

2018

2019

2021

139

Figure A.2 January geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four years of interest, mapped on a 2.5°
grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 30 geopotential metres
(gpm). Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown in shades of blue
and are also expressed in units of gpm. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020
reference period (IRI Climate and Society Map Room, n.d.).

925
hPa

500
hPa

250
hPa

Jan

2008

2018

2019

2021

140

Figure A.3 February geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four years of interest, mapped on a 2.5°
grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 30 geopotential metres
(gpm). Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown in shades of blue
and are also expressed in units of gpm. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020
reference period (IRI, n.d.).

925
hPa

500
hPa

250
hPa

Feb

2008

2018

2019

2021

Figure A.4 March geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four years of interest, mapped on a 2.5°
grid. Monthly geopotential height climatology values are shown as grey contours with a contour interval of 30 geopotential metres
(gpm). Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown in shades of blue
and are also expressed in units of gpm. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020
reference period (IRI, n.d.).
141

925
hPa

500
hPa

250
hPa

Mar

2008

2018

2019

2021

142

Figure A.5 April geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four years of interest, mapped on a 2.5° grid.
Monthly geopotential height climatology values are shown as grey contours with a contour interval of 30 geopotential metres (gpm).
Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown in shades of blue and are
also expressed in units of gpm. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference
period (IRI, n.d.).

925
hPa

500
hPa

250
hPa

Apr

2008

2018

2019

2021

Figure A.6 May geopotential height anomalies (at 250 hPa, 500 hPa, and 925 hPa) for the four years of interest, mapped on a 2.5° grid.
Monthly geopotential height climatology values are shown as grey contours with a contour interval of 30 geopotential metres (gpm).
Positive geopotential height anomalies are shown in yellow and orange, and negative anomalies are shown in shades of blue and are
also expressed in units of gpm. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset using the 1991 - 2020 reference
period (IRI, n.d.).
143

925
hPa

500
hPa

250
hPa

May

2018

2019

2021

144

Figure A.7 Contour maps of monthly (December, January, and February) average sea level pressure anomalies for the four water years
of interest, with respect to the 1991 - 2020 reference period. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset and
mapped in units of hectopascals (hPa), with a contour interval of 2 hPa. Blue lines represent 0 hPa. Original maps were modified to
highlight positive (yellow) and negative (red) anomalies (original image source: IRI, n.d.).

FEB

JAN

DEC

2008

A.2 Monthly Sea Level Pressure Anomaly

2018

2019

2021

145

Figure A.8 Contour maps of monthly (March, April, and May) average sea level pressure anomalies for the four years of interest, with
respect to the 1991 - 2020 reference period. Anomalies are calculated from the NCEP-NCAR Reanalysis dataset and mapped in units
of hectopascals (hPa), with a contour interval of 2 hPa. Blue lines represent 0 hPa. Original maps were modified to highlight positive
(yellow) and negative (red) anomalies (original image source: IRI, n.d.)

MAY

APR

MAR

2008