Editorial Type:
Article
Article Type:
Research Article

Patterns and Trend Analysis of Rain-on-Snow Events using Passive Microwave Satellite Data over the Canadian Arctic Archipelago Since 1987

Vincent Sasseville aCentre d’Applications et Recherches en Télédétection (Cartel), Université de Sherbrooke, Quebec, Quebec, Canada
bCentre d’Études Nordiques, Quebec, Quebec, Canada

Search for other papers by Vincent Sasseville in
Current site
Google Scholar
PubMed Close
https://orcid.org/0000-0002-2813-7178
,
Alexandre Langlois aCentre d’Applications et Recherches en Télédétection (Cartel), Université de Sherbrooke, Quebec, Quebec, Canada
bCentre d’Études Nordiques, Quebec, Quebec, Canada

Search for other papers by Alexandre Langlois in
Current site
Google Scholar
PubMed Close
,
Ludovic Brucker cNOAA/NESDIS/Center for Satellite Applications and Research, College Park, Maryland
dU.S. National Ice Center, Suitland, Maryland

Search for other papers by Ludovic Brucker in
Current site
Google Scholar
PubMed Close
, and
Cheryl Ann Johnson eWildlife Landscape Science Division, Environment and Climate Change Canada, Ottawa, Ontario, Canada

Search for other papers by Cheryl Ann Johnson in
Current site
Google Scholar
PubMed Close
Online Publication:
08 Feb 2024
Print Publication:
01 Feb 2024
DOI:
https://doi.org/10.1175/JHM-D-22-0218.1
Page(s):
311–324
Free access

Abstract

Climate change has a profound effect on Arctic meteorology extreme events, such as rain-on-snow (ROS), which affects surface state variable spatial and temporal variability. Passive microwave satellite images can help detect such events in polar regions where local meteorological and snow information is scarce. In this study, we use a detection algorithm using high-resolution passive microwave data to monitor spatial and temporal variability of ROS over the Canadian Arctic Archipelago from 1987 to 2019. The method is validated using data from several meteorological stations and atmospheric corrections have been applied to the passive microwave dataset. Our approach to detect ROS is based on two methods: 1) over a fixed time period (i.e., 1 November–31 May) throughout the study period and 2) using an a priori detection for snow presence before applying our ROS algorithm (i.e., length of studied winter varies yearly). Event occurrence is analyzed for each winter and separated by island groups of the Canadian Arctic Archipelago. Results show an increase in absolute ROS occurrence, mainly along the coasts, although no statistically significant trends are observed.

Significance Statement

Rain-on-snow (ROS) is known to have significant consequences on vegetation and fauna, especially widespread events. This study aimed to use a recent high-resolution dataset of passive microwave observations to investigate spatial and temporal trends in ROS occurrence in the Arctic. Results show that a global increase in event occurrence can be observed across the arctic.

© 2024 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Vincent Sasseville, vincent.sasseville@usherbrooke.ca

Abstract

Climate change has a profound effect on Arctic meteorology extreme events, such as rain-on-snow (ROS), which affects surface state variable spatial and temporal variability. Passive microwave satellite images can help detect such events in polar regions where local meteorological and snow information is scarce. In this study, we use a detection algorithm using high-resolution passive microwave data to monitor spatial and temporal variability of ROS over the Canadian Arctic Archipelago from 1987 to 2019. The method is validated using data from several meteorological stations and atmospheric corrections have been applied to the passive microwave dataset. Our approach to detect ROS is based on two methods: 1) over a fixed time period (i.e., 1 November–31 May) throughout the study period and 2) using an a priori detection for snow presence before applying our ROS algorithm (i.e., length of studied winter varies yearly). Event occurrence is analyzed for each winter and separated by island groups of the Canadian Arctic Archipelago. Results show an increase in absolute ROS occurrence, mainly along the coasts, although no statistically significant trends are observed.

Significance Statement

Rain-on-snow (ROS) is known to have significant consequences on vegetation and fauna, especially widespread events. This study aimed to use a recent high-resolution dataset of passive microwave observations to investigate spatial and temporal trends in ROS occurrence in the Arctic. Results show that a global increase in event occurrence can be observed across the arctic.

© 2024 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Vincent Sasseville, vincent.sasseville@usherbrooke.ca
Keywords: Arctic; Climate change; Microwave observations; Remote sensing

1. Introduction

Over the past four decades, the Arctic underwent the most climate variability and changes as a response to strong climate-related positive feedback (Serreze and Barry 2011). The Arctic has warmed nearly 4 times faster than the rest of the Earth since 1979 (Rantanen et al. 2022). Patterns of negative anomalies in snow mass balances and snow cover spatial extent have been observed, significantly affecting how the cryosphere responds to climate change (Cullather et al. 2016; Liston and Hiemstra 2011; Derksen and Brown 2012; Larue et al. 2017). Furthermore, recent model simulations predict an increase in both the mean and the variability of arctic precipitations over the next decades (Bintanja et al. 2020) that would become more frequent (Song and Liu 2017). Satellite-based studies have shown that rain-on-snow (ROS) occurrences are increasing as a response to this accelerated warming (Rennert et al. 2009; Langlois et al. 2017; Dolant et al. 2018a). However, their cumulative impacts on snow and the environment remain unknown.

ROS events and increasingly warming temperatures lead to the presence of liquid water in the snow cover that alters energy transfer mechanisms (Rennert et al. 2009; Weismüller et al. 2011; Riseborough et al. 2008), causing uncertainties in climate projections. Water in the snowpack creates ice crusts and melt–freeze layers that can have negative impacts on foraging conditions for the Arctic’s ungulates such as caribou (Rangifer tarandus) and muskoxen (Ovibos moschatus) (Putkonen et al. 2009; Langlois et al. 2017; Sokolov et al. 2016). Long-term remote sensing–based approaches allow quantifying spatial and temporal changes in the melt–freeze cycle that results from ROS (e.g., Marmy et al. 2013)

The spatial and temporal occurrence variability of ROS events remains widely unknown, despite their impacts on the surface energy balance and snow condition, because of a sustained lack of data. The lack of weather stations in the Arctic makes in situ monitoring difficult, especially regarding ROS events spatial patterns and trends. Liston and Hiemstra (2011) used snow modeling alongside atmospheric forcing data to analyze temporal trends and the results suggested an increase in ROS days of +0.03 days decade−1 between 1979 and 2009. Passive microwaves have also been used as a method to monitor ROS (Dolant et al. 2016; Grenfell and Putkonen 2008) given the high sensitivity of microwave brightness temperatures (Tb) to wet snow (Comiso et al. 2003). The water changes the snow properties and its microwave dielectric contrast before (i.e., dry) and after (i.e., wet) an ROS event. In normal and dry conditions, the values of Tb at 19 GHz are higher than Tb at 37 GHz given the enhanced scattering loss at high frequencies and shorter wavelength but the opposite is observed in wet conditions. Unfortunately, the coarse spatial resolution of passive microwaves (10–25 km) leads to biases in errors of omissions and commissions in existing algorithms (Dolant et al. 2018a). A new passive microwave dataset [NASA’s Making Earth System Data Records for Use in Research Environments (MEaSUREs)] has been published with a gridded resolution of 3.125 km × 3.125 km for Tb values at 37 GHz and 6.25 km × 6.25 km at 19 GHz (Brodzik et al. 2016). The 3.125 km × 3.125 km grid is used in this study to get a better precision for ROS event detection.

The main objective of this paper is to produce an ROS occurrence database over the Canadian Arctic Archipelago (CAA) from 1979 to the present using an updated detection algorithm (Dolant et al. 2016, 2018a). The objectives consist of 1) producing ROS occurrence anomaly maps to share with the wide scientific community and 2) evaluating the spatial and temporal variability in event occurrence across the CAA.

2. Data and method

a. Study site

The algorithm is run over all the islands of the CAA (Fig. 1). Baffin Island was excluded because no caribou live there. We divided the area into seven island groups as follows:

  • Victoria Island (Victoria);

  • Banks Island (Banks);

  • Island group: Prince of Wales, Somerset, Russell, and Boothia Peninsula (Boothia)

  • Island group: Axel Heiberg, Ellesmere, Ellef Ringnes, Corwall, Graham, Meighen, Coburg, and Amund (Axel and Ellesmere);

  • Island group: Melville, Prince Patrick, Eglinton, Emerald, and Byam Martin (Melville);

  • Island group: Mackenzie, Brock, and Borden (Mackenzie);

  • Island group: Bathurst, Cornwallis, Little Cornwallis, Lougheed, Devon, and Baillie-Hamilton (BIC).

Fig. 1.
Fig. 1.

Area covered by the seven island groups studied in this report.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

This list only cites the biggest islands in each group as there are about 90 major islands. The total area covers approximately 1 400 000 km2.

b. Passive microwave satellite data

1) MEaSUREs

Passive microwave data were extracted from the Calibrated Enhanced-Resolution Passive Microwave Daily EASE-Grid 2.0 Brightness Temperature ESDR, Version 1 (MEaSUREs) dataset (Brodzik et al. 2016). The dataset consists of gridded passive microwave brightness temperature (Tb) data from the following instruments:

  • Scanning Multichannel Microwave Radiometer (SMMR) on Nimbus-7;

  • Special Sensor Microwave Imager (SSM/I) on DMSP 5D-2/F8, F10, F11, F13, F14, DMSP 5D-3/F15;

  • Special Sensor Microwave Imager/Sounder (SSMIS) on DMSP 5D-3/F16, F17, F18, F19;

  • Advanced Microwave Scanning Radiometer–Earth Observing System (AMSR-E) on Aqua.

Table 1 presents a selection of sensors and periods. Brodzik et al. (2016) used an overlap technique to enhance the resolution up to 3.125 km × 3.125 km. Two frequencies were used to detect ROS (see section 2c): 19 and 37 GHz with both horizontal and vertical polarizations. As such, the MEaSUREs database provides grids with a resolution of 6.25 km × 6.25 km for the 19-GHz Tb dataset and grids with a resolution of 3.125 km × 3.125 km for the 37-GHz Tb dataset. We divided by 4 the pixels from the 19-GHz dataset to fit the 37-GHz grids and created a temporal dataset from 1 July 1987 to 30 June 2018. We excluded data prior to September 1987 because record gaps made unavailable a high number of data points, making it too scarce to meet the objectives highlighted above. The values from the twice daily temporal resolution were averaged to give one daily Tb observation at 3.125 km for each frequency and polarization (i.e., 4 Tb values daily) starting using the data from the SSM/I F08 sensor (Table 1). Data with a 37-GHz frequency with vertical polarization from 1 August 2016 to the end of the period were retrieved from the F16 SSMIS sensor instead of the F17, which the dataset provider flagged as too low quality (Brodzik et al. 2016). In this study, each pixel is tested with the algorithm each day, therefore, the numbers in the results do not represent the number of events since one event can affect multiple pixels for multiple days. The GRP [Eq. (4)] response to water on the surface of the snowpack can linger for multiple days after an event or have a different cause, inflating the numbers again.

Table 1.

List of sensors selected to form the database used for ROS detection.
Table 1.

2) The North American Regional Reanalysis and atmospheric corrections

We applied an atmospheric correction to the Tb dataset using the North American Regional Reanalysis (NOAA/ESRL/Physical Sciences Laboratory 2020). More specifically, the total precipitable water (pwat) for the entire atmospheric column was extracted to derive the atmospheric contributions to Tb and to the atmosphere’s transmissivity. The NARR database temporal period overlaps with MEaSUREs, i.e., beginning before 1987. The product has a resolution of 0.3°, which consists of approximately 32 km × 32 km depending on latitude. For each pixel of the MEaSUREs database, the closest NARR pixels are used to obtain a value of pwat to use for the correction; this approach has been used in peer-reviewed articles for over a decade (Roy 2014). NARR is a commonly used dataset for meteorological data over the Canadian Arctic, and it has been coupled with microwave datasets in many studies before (Montpetit et al. 2013; Dupont et al. 2014; Picard et al. 2013; Roy 2014). The corrected measurements (Tbcorr) were calculated using Eq. (1). The correction method uses the Millimeter-Wave Propagation Model, allowing the retrieval of both transmissivity and atmospheric Tb (Dolant et al. 2018b):
Tbcorr=Tbmes−at↑τ,
where Tbmes is the Tb provided by the MEaSUREs database, at↑ is the upward contribution of the atmosphere, which is subtracted from the final measurement, τ is the transmissivity of the atmosphere, and at↑ and τ are both calculated using a linear function (α and β) of the precipitable water (pwat) of the atmosphere [Eq. (2)]. The values α and β come from Roy (2014) (Table 2):
at↑,τ=α×pwat+β.
Table 2.

Linear functions of pwat used to derive the atmospheric brightness temperature contribution and transmissivity.
Table 2.

c. ROS detection algorithm

Dolant et al. (2016, 2018a) developed the original passive microwave detection algorithm for ROS. During an ROS event, liquid water accumulates at the snow surface, which is typically very dense in the Arctic, where the snowpack consists mainly of a wind slab on top of a hoar layer (Royer et al. 2021). The water changes the snow properties and its microwave dielectric contrast before (i.e., dry) and after (i.e., wet) a, ROS event. In normal and dry conditions, the values of Tb at 19 GHz are higher than Tb at 37 GHz given the enhance scattering loss at high frequencies and shorter wavelength. A snow surface wet from rainwater is warmer than the underlying snowpack and leads to microwave emission increasing, proportionally to the frequency. The increase is higher for horizontal polarization given the horizontal nature of the hard snow surface.

As such, the detection algorithm first uses the gradient ratio (GR) from Tb at 19 GHz and Tb at 37 GHz using Eq. (3):
GR(37,19)p=Tb(p,37GHz)−Tb(p,19GHz)Tb(p,37GHz)+Tb(p,19GHz),
where Tb(p, f) is the brightness temperature of the frequency f and the polarization p.
Using both GRυ and GRh we could define the GRP (Dolant et al. 2016, 2018a) using Eq. (4):
GRP(37,19)=GR(37,19)υGR(37,19)h.
The GRP is a good indicator of how severe the event is where negative values indicate the Tb reversal, i.e., Tb at 37 GHz > Tb at 19 GHz on which we could apply a threshold after an omission and commission analysis on detection accuracy. The GRP was set as −10 in Dolant et al. (2018a) using visual observations of precipitations in 18 meteorological stations of the Arctic. In this study, we conducted the threshold assessment using precipitation measurements from meteorological stations rather than visual precipitation type observations which can be biased by the observer. This method comes with uncertainties as liquid water can form on top of a snowpack for different reasons, such as warm sunny days. If only severe ROS events were relevant, we could lower the GRP threshold even more at the cost of increasing omissions of general events. We revaluated the GRP threshold in the next section with the same goal of finding the middle ground between commission and omission.

d. Meteorological stations and precipitations measurements

The Environment and Climate Change Canada (ECCC) station network recorded the meteorological data used for the GRP threshold assessment in this study and includes stations in the CAA. We selected stations covering the 1987–2018 time period measuring daily accumulated rain precipitations (automated precipitation gauges) and snow depth. The following six stations were selected: Cambridge Bay (69.05°N, 79.02°W), Sachs Harbour (72.00°N, 125.27°W), Resolute Bay (74.72°N, 94.97°W), Gjoa Haven (68.64°N, 95.85°W), Taloyoak (69.55°N, 93.58°W), and Ulukhaktok (70.76°N, 117.81°W). The chosen stations cover many longitudes while being approximately on the same latitude. Rain and snow depth data were automatically extracted from each station over the 1987–2018 period to drive our ROS events count with the approach described below. An ROS was counted if the data showed at least 3 mm of rain in combination with at least 10 cm of snow height for a day. If an ROS was detected using our algorithm but was not measured at the station, we counted it as a commission. We also counted an omission when our algorithm did not detect an ROS observed by the station. We calculated the error percentage by adding the total commission and omission and dividing it by the total ROS detected (ROStot) by the meteorological stations:
error=100×[(commission+omission)/ROStot].
The error can be calculated over many GRP thresholds to evaluate the optimized threshold representing the best compromise between omissions and commissions.

e. ROS count method

We applied the GRP algorithm to the daily corrected (atmospheric corrections) Tb measurements. If an ROS event was detected on a pixel, the value 1 was attributed; otherwise, the value was 0. We applied the algorithm on all inland pixels of an island group, except the pixels next to the water. Then, using two approaches detailed in the next paragraph, we accumulated the occurrences per winter by adding the daily values from each pixel to create daily occurrence maps, yearly accumulation maps, and yearly anomalies maps. We obtained anomalies by calculating the winter average ROS occurrence for each pixel over the period, then by calculating the difference between that average and the number of occurrences for a particular winter.

The GRP algorithm was applied following two “temporal” approaches: 1) “fixed period” and 2) “snow condition.” The fixed period applies the GRP algorithm over a fixed duration. We defined winter from 1 November to 31 May, providing a conservative estimate of the snow cover duration. This approach allows a constant “observation period,” which is useful when monitoring temporal changes in ROS occurrences over the years. In contrast, the snow condition approach uses a snow detection algorithm to detect the presence of a snow cover, then applies the GRP algorithm. The snow presence is detected when Tb at 19 GHz is higher than the Tb at 37 GHz in the vertical polarization, which is a well-used criteria in several snow water equivalent or snow depth microwave algorithms (Shi et al. 2016; Foster et al. 1984; Armstrong and Brodzik 2002). This approach makes temporal trends hard to analyze given the observational period changes over time but provides a better assessment of spatial patterns during winter by including the entire season as opposed to a fixed period.

3. Results

a. GRP threshold assessment

Dolant et al. (2016, 2018a) suggested a threshold between −2 and −10 based on errors of omission versus commission. Using visual observations of precipitation types, they suggested reducing the threshold to account only for major events impacting caribou foraging conditions (small events do not affect foraging conditions). In our case, rather than using visual observations, which are scarce, we used in situ precipitation measurements from the stations equipped with precipitation gauges (listed in section 2d). For data between 1987 and 2018, we applied the GRP over the pixels in which the meteorological stations are located to quantify the algorithm precision over a range of GRP thresholds. Results from Fig. 2 highlight the evolution of commissions (Fig. 2a), omissions (Fig. 2b), and the total error [Eq. (6); Fig. 2c] as a function of the GRP threshold:
100×correct detectioncommission+omission.
Figure 2 suggests that the total error significantly increases above a threshold from −10 to −5. Much weaker ROS events were detected when GRP was close to zero, increasing the likelihood of commission errors compared to smaller GRP values that better represent a significant ROS event. Interestingly, Dolant et al. (2018a) suggested a threshold of −10 as a compromise between catching main events while keeping the total error low. Results from Fig. 2 suggest that a threshold between −5 and −15 would provide a total error ranging between 1% and 2.5%. Thus, our results provide strong quantitative support for the use of the −10 threshold to identify ROS events reported by Dolant et al. (2018a). This is in strong agreement with Dolant et al. (2018a), who used a different meteorological data source and analysis, thus confirming the choice of a threshold of −10 in this study to produce the ROS maps.
Fig. 2.
Fig. 2.

Overview of the algorithm validation results using data from meteorological stations. (a) Number of commissions in relation to the GRP threshold. (b) Number of omissions in relation to the GRP [Eq. (4)] threshold. (c) Error in percent [Eq. (5)] in relation to the GRP threshold.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

b. ROS anomalies

To compute ROS occurrence anomalies, we used a threshold value of −10 on the GRP and computed yearly anomalies following the fixed period and the snow condition approaches for each island group. We expected ROS occurrences to be higher in this second set of maps because the winter monitoring period spans over all months with snow cover.

1) Fixed period

The fixed period method ensures the number of days examined each year for ROS events remains constant. Figure 3 provides a broad idea of the region that received more ROS on average. It illustrates the number of ROS occurrences per year averaged for the 32 years from 1987 to 2018 included. The algorithm is not valid over glaciers, and a land mask has been used to exclude them. The higher occurrences on Ellesmere Island are likely false positives caused by the proximity of the glacier. Other regions with higher occurrences are the coastlines of the southern islands.

Fig. 3.
Fig. 3.

Annual average ROS occurrences from 1987 to 2019 with a fixed winter season from 1 Nov to 31 May as a scale of gray maxed at 4 ROS.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Figures 4 and 5 are examples chosen arbitrarily of anomalies of ROS occurrences detected across the CAA for, respectively, the 2010/11 and 2015/16 seasons. We calculated the anomaly with respect to the means of ROS detected per year and per pixel over the 1987–2018 period. Often, the ROS observations were along the coasts despite excluding pixels touching the water. Meanwhile, the ROS events detected were much closer to the average toward the middle of the islands.

Fig. 4.
Fig. 4.

Rain-on-snow occurrence anomalies map for winter 2010/11 using the fixed period method.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Fig. 5.
Fig. 5.

Rain-on-snow occurrence anomalies map for winter 2015/16 using the fixed period method.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Figure 6 shows the annual ROS anomalies averaged per island group. Values are color-coded from low (blue) to high (orange). Different regions show different evolutions. For instance, Mackenzie Island group’s ROS occurrence has not changed much over the years despite a large and intense ROS season during the winter of 2002/03. In contrast, BIC demonstrates high variability in ROS occurrence when compared to other island groups. We observed a general peak in ROS occurrence across the CAA for winter 2014/15, but then fewer occurrences happened in recent years. Despite high interannual variability in ROS anomalies, Table 3 suggests a current increase in positive anomalies compared to past conditions for Banks, Boothia, Mackenzie, and Melville.

Fig. 6.
Fig. 6.

Spatial average of ROS occurrence anomalies with the fixed period method. Blue represents a negative anomaly and orange represents a positive anomaly compared to the 1987–2018 average for each island group.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Table 3.

Percentage of years with positive average anomalies using the fixed period method.
Table 3.

2) Snow condition period

The snow condition method produced similar results to the fixed period method. Figure 7 highlights the average number of events each year. By using the snow condition method, we expanded the monitored period, thus inevitably increasing the total of occurrences detected. This did not seem to result in any major changes in the spatial or temporal patterns of anomalies. Figure 7 highlights areas where the major changes in ROS occurred across the CAA. With a longer observation period, more ROS events were detected, especially on the northernmost islands, since more events are more likely to occur at the beginning and the end of the snow season where there is a mix of rain and snow. With the fixed period method, we likely missed this moment for the northern islands. Once again, we observed more average occurrences along coastlines, but also across every island. Figures 8 and 9 show examples of anomaly for the years 2010 and 2015, respectively. Once again, stronger anomalies were observed along the coastlines and revealed the location where the meteorological systems produced more ROS for each year.

Fig. 7.
Fig. 7.

Annual average ROS occurrences from 1987 to 2019 with the snow condition method as a scale of gray maxed at 4 ROS.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Fig. 8.
Fig. 8.

Rain-on-snow occurrence anomalies map for winter 2010/11 using the snow condition method.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Fig. 9.
Fig. 9.

Rain-on-snow occurrence anomalies map for winter 2015/16 using the snow condition method.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Figure 10 highlights the temporal evolution of the averaged island yearly anomalies, using the snow condition approach. The behavior in trend occurrence across the islands is different than with the fixed period method (Table 4). Only 1 out of 7 island complexes have more ROS occurrence observed in the latter period compared to 4 out of 7 for the fixed period method. This suggests that the shortening of the snow season might lead to less ROS events observed with time.

Fig. 10.
Fig. 10.

Spatial average of ROS occurrence anomalies with the snow detection method. Blue refers to a negative anomaly and orange refers to a positive anomaly compared to the 1987–2018 average for each island group.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Table 4.

Percentage of years with positive average anomalies using the snow condition method.
Table 4.

The ROS events detected for every island varied widely. While in absolute more ROS events were detected in the latter part of the study period, we detected no significant trend. All the data are available on the https://grimp.ca/ server as a GeoTIFF raster map. The database contains rasters showing the period total yearly occurrences and yearly anomalies for both the snow condition and fixed period methods as shown in Table 5.

Table 5.

Number of rasters produced covering the entire study area.
Table 5.

4. Discussion

We selected the algorithm threshold based on data from various meteorological stations in the CAA, besides the recommendations of previous works. The validation at the meteorological stations confirms the algorithm is working, which was already demonstrated in Dolant et al. (2018a) and Langlois et al. (2017). The method does yield an number of omissions and commissions. Events other than ROS can cause liquid on the surface of the snowpack and trigger the algorithm (e.g., melt), or some ROS events might be too light to be detected (e.g., light drizzle). The threshold can be adjusted based on the study objectives. The biases of the data are unknown and remain a source of uncertainty in this study. The rasters were produced at a resolution of 3.125 km × 3.125 km since we do not expect any significant change in the Tb at 19 GHz given the homogeneous nature of tundra and polar desert snow (Royer et al. 2021), and we then optimize the information for the Tb at 37 GHz, which causes the reversal (Dolant et al. 2016) given its increased sensitivity to wetness. Using the raster products, we could assess spatial and temporal changes in ROS occurrence across the CAA. Our results show a high variability in occurrence anomalies between different island groups. On large islands, positive anomalies are observed along the coastline, with few events in the mainland. Topography likely does not impact GRP measurements but creates orographic effects that affect ROS occurrences. For the fixed period method, most island groups experienced an increase in “positive years,” where anomalies were especially strong toward the end of the study period, especially for the 2015/16 season, which was the warmest over the studied period (Cullather et al. 2016). When considering snow on the ground, we observe less ROS in the later period of the study. As highlighted by Moon et al. (2021) negative anomaly of the snow cover extent is observed in the latter period of our study. This is especially true for the spring season (Derksen and Mudryk 2023) where most ROS can be expected (Dolant et al. 2018a). The lower sea ice coverage during warmer periods might explain the results when open water can increase ROS probability. We conducted a preliminary assessment of sea ice concentration surrounding the island groups with the sea ice coverage information from the work of Couture (2022), who computed ice concentration maps from the Canadian Ice Service Digital Archive (CISDA) for the eastern Arctic and western Arctic. Ice forecasters interpret all available ice information from 1978 onward to create ice charts in the CISDA and to map the different ice types and concentration areas. From this raster dataset, we calculated the average open water area extent in a buffer of 50 km around Victoria Island and Banks Island. For each month, we calculated and compared the total average area of open water in the buffer with the number of ROS events detected for that month. Figure 11 shows examples for the month of October for both islands. We used the snow condition approach for the ROS detection to ensure false positives were excluded where a detection would occur when there was no snow on the ground.

Fig. 11.
Fig. 11.

Number of ROS events detected compared to the total area of open water around a buffer of 50 km from the island in October for each year. (a) Victoria Island. (b) Banks Island.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Since the islands are relatively wide, we compared ROS occurrences with open water on a finer scale. We thus divided the island and the ice buffer in four quadrants: southwest, southeast, northeast, and northwest (Fig. 12).

Fig. 12.
Fig. 12.

Number of ROS events detected compared to the total area of open water around a buffer of 50 km from Victoria Island in October for each year: (a) northwest, (b) northeast, (c) southwest, and (d) southeast.

Citation: Journal of Hydrometeorology 25, 2; 10.1175/JHM-D-22-0218.1

Although preliminary results suggest that an increase in ROS can be attributed to an increase in open water extent, the linear regressions are not statistically significant, suggesting that we need further work to find a better explanation. Many factors can influence ROS occurrences including wind, temperature, and humidity. Links with local ice conditions might not be trivial.

5. Conclusions

Using the MEaSUREs brightness temperature dataset with an atmospheric correction, we applied an ROS occurrence algorithm over daily CAA data for the period extending from fall 1987 to spring 2018 at a 3.125-km spatial resolution. The number of occurrences, along with yearly anomalies, was investigated using two methods based on how snow presence was considered. Following the main objective, we created a database of raster products on ROS occurrence that can now be used for further research. The dataset contains 128 raster maps including total ROS occurrences and anomalies (32 years each, for both fixed and snow condition approaches) and can be updated along with the updates of the MEaSUREs dataset. The spatial trend analysis indicated an enhanced detection of positive anomalies in the second half of the study period, with most detections expectedly occurring along coastlines, given the proximity of open water. When using a fixed period, the southernmost islands showed a higher number of occurrences, which was not the case when using the snow condition approach. Such results suggest that for the northern parts of the CAA, ROS events tended to occur earlier and later in the season when they were not counted using the fixed method. As for the temporal trend analysis, we observed a strong variability with more ROS occurrences in the later year of the period when using the fixed period winter season but observed less ROS when counting for the presence of snow. It suggests the shortening of the snow season period might lead to less ROS in general. It should be noted that the threshold for ROS detection of GRP = −10 had been validated with data from Arctic snow and should be limited to this region. Different regions should use this method with a revaluated threshold.

This database can be used to investigate the causal relationship with other variables and processes. The hypothesis that ROS occurrences are linked to the amount of free water around a region should be investigated further for all islands because our preliminary assessment of Victoria Island and Banks Island showed that some correlation during the fall might exist when ice coverage varies more, but we found no statistically significant correlation. Future studies could compare the method described here with other methods of ROS detection to better understand the limits of different approaches. This new dataset can also be used as input in ecological models, such as that in the works presented in Gautier (2022) and Martineau (2020) that used snow model outputs and an explanatory variable in ecological models to predict the presence of Peary caribou populations. It would be interesting to combine ROS rasters with their approach aiming at improving our ability to predict the animal presence and expand the approach to other ungulates of the Arctic. Finally, we hope this work can lead to an improved understanding of the processes controlling ROS occurrence and how such events affect other climate variables, improving our understanding of the Arctic’s climate system.

Acknowledgments.

This work was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC), Environment and Climate Change Canada (ECCC), and the Fonds de recherche du Québec–Nature et Technologies (FRQNT). The scientific results and conclusions, as well as any views or opinions expressed herein, are those of the authors and do not necessarily reflect those of NOAA or the U.S. Department of Commerce.

Data availability statement.

The authors were unable to find a valid data repository for the data produced in this study. These data are available from vincent.sasseville@usherbrooke.ca at the University of Sherbrooke.

REFERENCES

  • Armstrong, R. L., and M. J. Brodzik, 2002: Hemispheric-scale comparison and evaluation of passive-microwave snow algorithms. Ann. Glaciol., 34, 38–44, https://doi.org/10.3189/172756402781817428.

    • Search Google Scholar
    • Export Citation

  • Bintanja, R., K. van der Wiel, E. C. van der Linden, J. Reusen, L. Bogerd, F. Krikken, and F. M. Selten, 2020: Strong future increases in Arctic precipitation variability linked to poleward moisture transport. Sci. Adv., 6, eaax6869, https://doi.org/10.1126/sciadv.aax6869.

    • Search Google Scholar
    • Export Citation

  • Brodzik, M. J., D. G. Long, M. A. Hardman, A. Paget, and R. Armstrong, 2016: MEaSUREs calibrated enhanced-resolution passive microwave daily ease-grid 2.0 brightness temperature ESDR, version 1. NASA National Snow and Ice Data Center Distributed Active Archive Center, accessed 22 February 2020, https://doi.org/10.5067/measures/cryosphere/nsidc-0630.001.

  • Comiso, J. C., J. Yang, S. Honjo, and R. A. Krishfield, 2003: Detection of change in the Arctic using satellite and in situ data. J. Geophys. Res., 108, 3384, https://doi.org/10.1029/2002JC001347.

    • Search Google Scholar
    • Export Citation

  • Couture, G., 2022: Analyses spatiotemporelles des conditions de glace de mer et des tendances de formation des polynies de l’Archipel Arctique Canadien. M.S. thesis, Département de géomatique appliquée Faculté des lettres et sciences humaines, Université de Sherbrooke, 89 pp., https://savoirs.usherbrooke.ca/handle/11143/19074.

  • Cullather, R. I., Y.-K. Lim, L. N. Boisvert, L. Brucker, J. N. Lee, and S. M. Nowicki, 2016: Analysis of the warmest Arctic winter, 2015–2016. Geophys. Res. Lett., 43, 10 808–10 816, https://doi.org/10.1002/2016GL071228.

    • Search Google Scholar
    • Export Citation

  • Derksen, C., and R. Brown, 2012: Spring snow cover extent reductions in the 2008–2012 period exceeding climate model projections. Geophys. Res. Lett., 39, L19504, https://doi.org/10.1029/2012GL053387.

    • Search Google Scholar
    • Export Citation

  • Derksen, C., and L. Mudryk, 2023: Assessment of Arctic seasonal snow cover rates of change. Cryosphere, 17, 1431–1443, https://doi.org/10.5194/tc-17-1431-2023.

    • Search Google Scholar
    • Export Citation

  • Dolant, C., A. Langlois, B. Montpetit, L. Brucker, A. Roy, and A. Royer, 2016: Development of a rain-on-snow detection algorithm using passive microwave radiometry. Hydrol. Processes, 30, 3184–3196, https://doi.org/10.1002/hyp.10828.

    • Search Google Scholar
    • Export Citation

  • Dolant, C., A. Langlois, L. Brucker, A. Royer, A. Roy, and B. Montpetit, 2018a: Meteorological inventory of rain-on-snow events in the Canadian Arctic Archipelago and satellite detection assessment using passive microwave data. Phys. Geogr., 39, 428–444,https://doi.org/10.1080/02723646.2017.1400339.

    • Search Google Scholar
    • Export Citation

  • Dolant, C., B. Montpetit, A. Langlois, L. Brucker, O. Zolina, C. A. Johnson, A. Royer, and P. Smith, 2018b: Assessment of the barren ground Caribou die-off during winter 2015–2016 using passive microwave observations. Geophys. Res. Lett., 45, 4908–4916, https://doi.org/10.1029/2017GL076752.

    • Search Google Scholar
    • Export Citation

  • Dupont, F., G. Picard, A. Royer, M. Fily, A. Roy, A. Langlois, and N. Champollion, 2014: Modeling the microwave emission of bubbly ice: Applications to blue ice and superimposed ice in the Antarctic and Arctic. IEEE Trans. Geosci. Remote Sens., 52, 6639–6651, https://doi.org/10.1109/TGRS.2014.2299829.

    • Search Google Scholar
    • Export Citation

  • Foster, J. L., D. K. Hall, A. T. C. Chang, and A. Rango, 1984: An overview of passive microwave snow research and results. Rev. Geophys., 22, 195–208, https://doi.org/10.1029/RG022i002p00195.

    • Search Google Scholar
    • Export Citation

  • Gautier, C., 2022: Étude de la qualité d’habitat et des patrons migratoires du caribou de Peary (Rangifer tarandus pearyi) à l’aide d’un modèle thermodynamique de neige, des anomalies de glace de mer et du savoir traditionnel. M.S. thesis, Dépt. de géomatique appliquée Faculté des lettres et sciences humaines, Université de Sherbrooke, 155 pp., https://savoirs.usherbrooke.ca/handle/11143/18415.

  • Grenfell, T. C., and J. Putkonen, 2008: A method for the detection of the severe rain-on-snow event on Banks Island, October 2003, using passive microwave remote sensing. Water Resour. Res., 44, W03425, https://doi.org/10.1029/2007WR005929.

    • Search Google Scholar
    • Export Citation

  • Langlois, A., and Coauthors, 2017: Detection of rain-on-snow (ROS) events and ice layer formation using passive microwave radiometry: A context for Peary caribou habitat in the Canadian Arctic. Remote Sens. Environ., 189, 84–95, https://doi.org/10.1016/j.rse.2016.11.006.

    • Search Google Scholar
    • Export Citation

  • Larue, F., A. Royer, D. De Sève, A. Langlois, A. Roy, and L. Brucker, 2017: Validation of GlobSnow-2 snow water equivalent over eastern Canada. Remote Sens. Environ., 194, 264–277, https://doi.org/10.1016/j.rse.2017.03.027.

    • Search Google Scholar
    • Export Citation

  • Liston, G. E., and C. A. Hiemstra, 2011: The changing cryosphere: Pan-Arctic snow trends (1979–2009). J. Climate, 24, 5691–5712, https://doi.org/10.1175/JCLI-D-11-00081.1.

    • Search Google Scholar
    • Export Citation

  • Marmy, A., N. Salzmann, M. Scherler, and C. Hauck, 2013: Permafrost model sensitivity to seasonal climatic changes and extreme events in mountainous regions. Environ. Res. Lett., 8, 035048, https://doi.org/10.1088/1748-9326/8/3/035048.

    • Search Google Scholar
    • Export Citation

  • Martineau, C., 2020: Couplage du logiciel de modélisation de l’habitat MaxEnt à des simulations du couvert nival pour l’amélioration de la prédiction de présence du caribou de Peary. M.S. thesis, Département de géomatique appliquée Faculté des lettres et sciences humaines, Université de Sherbrooke, 92 pp., https://savoirs.usherbrooke.ca/handle/11143/17006.

  • Montpetit, B., A. Royer, A. Roy, A. Langlois, and C. Derksen, 2013: Snow microwave emission modeling of ice lenses within a snowpack using the microwave emission model for layered snowpacks. IEEE Trans. Geosci. Remote Sens., 51, 4705–4717, https://doi.org/10.1109/TGRS.2013.2250509.

    • Search Google Scholar
    • Export Citation

  • Moon, T. A., M. L. Druckenmiller, and R. L. Thoman, 2021: Arctic report card 2021: Executive summary. NOAA Tech. Rep. OAR ARC 21-01, 4 pp., https://doi.org/10.25923/5s0f-5163.

  • NOAA/ESRL/Physical Sciences Laboratory, 2020: NCEP North American regional reanalysis. NOAA, accessed 10 November 2020, https://psl.noaa.gov/data/gridded/data.narr.html.

  • Picard, G., L. Brucker, A. Roy, F. Dupont, M. Fily, A. Royer, and C. Harlow, 2013: Simulation of the microwave emission of multi-layered snowpacks using the Dense Media Radiative Transfer theory: The DMRT-ML model. Geosci. Model Dev., 6, 1061–1078, https://doi.org/10.5194/gmd-6-1061-2013.

    • Search Google Scholar
    • Export Citation

  • Putkonen, J., T. C. Grenfell, K. Rennert, C. Bitz, P. Jacobson, and D. Russell, 2009: Rain on snow: Little understood killer in the North. Eos, Trans. Amer. Geophys. Union, 90, 221–222, https://doi.org/10.1029/2009EO260002.

    • Search Google Scholar
    • Export Citation

  • Rantanen, M., A. Y. Karpechko, A. Lipponen, K. Nordling, O. Hyvärinen, K. Ruosteenoja, T. Vihma, and A. Laaksonen, 2022: The Arctic has warmed nearly four times faster than the globe since 1979. Commun. Earth Environ., 3, 168, https://doi.org/10.1038/s43247-022-00498-3.

    • Search Google Scholar
    • Export Citation

  • Rennert, K. J., G. Roe, J. Putkonen, and C. M. Bitz, 2009: Soil thermal and ecological impacts of rain on snow events in the circumpolar Arctic. J. Climate, 22, 2302–2315, https://doi.org/10.1175/2008JCLI2117.1.

    • Search Google Scholar
    • Export Citation

  • Riseborough, D., N. Shiklomanov, B. Etzelmüller, S. Gruber, and S. Marchenko, 2008: Recent advances in permafrost modelling. Permafrost Periglacial Processes, 19, 137–156, https://doi.org/10.1002/ppp.615.

    • Search Google Scholar
    • Export Citation

  • Roy, A., 2014: Modélisation de l’émission micro-onde hivernale en forêt boréale Canadienne. Ph.D. thesis, Université de Sherbrooke, 242 pp., http://savoirs.usherbrooke.ca/handle/11143/70.

  • Royer, A., F. Domine, A. Roy, A. Langlois, N. Marchand, and G. Davesne, 2021: New northern snowpack classification linked to vegetation cover on a latitudinal mega-transect across northeastern Canada. Écoscience, 28, 225–242, https://doi.org/10.1080/11956860.2021.1898775.

    • Search Google Scholar
    • Export Citation

  • Serreze, M. C., and R. G. Barry, 2011: Processes and impacts of Arctic amplification: A research synthesis. Global Planet. Change, 77, 85–96, https://doi.org/10.1016/j.gloplacha.2011.03.004.

    • Search Google Scholar
    • Export Citation

  • Shi, J., C. Xiong, and L. Jiang, 2016: Review of snow water equivalent microwave remote sensing. Sci. China Earth Sci., 59, 731–745, https://doi.org/10.1007/s11430-015-5225-0.

    • Search Google Scholar
    • Export Citation

  • Sokolov, A. A., N. A. Sokolova, R. A. Ims, L. Brucker, and D. Ehrich, 2016: Emergent rainy winter warm spells may promote boreal predator expansion into the Arctic. Arctic, 69, 121–129, https://doi.org/10.14430/arctic4559.

    • Search Google Scholar
    • Export Citation

  • Song, M., and J. Liu, 2017: The role of diminishing Arctic sea ice in increased winter snowfall over northern high-latitude continents in a warming environment. Acta Oceanol. Sin., 36, 34–41, https://doi.org/10.1007/s13131-017-1021-3.

    • Search Google Scholar
    • Export Citation

  • Weismüller, J., U. Wollschläger, J. Boike, X. Pan, Q. Yu, and K. Roth, 2011: Modeling the thermal dynamics of the active layer at two contrasting permafrost sites on Svalbard and on the Tibetan Plateau. Cryosphere, 5, 741–757, https://doi.org/10.5194/tc-5-741-2011.

    • Search Google Scholar
    • Export Citation
Save
Cover Journal of Hydrometeorology
Journal of Hydrometeorology

  • Armstrong, R. L., and M. J. Brodzik, 2002: Hemispheric-scale comparison and evaluation of passive-microwave snow algorithms. Ann. Glaciol., 34, 38–44, https://doi.org/10.3189/172756402781817428.

    • Search Google Scholar
    • Export Citation

  • Bintanja, R., K. van der Wiel, E. C. van der Linden, J. Reusen, L. Bogerd, F. Krikken, and F. M. Selten, 2020: Strong future increases in Arctic precipitation variability linked to poleward moisture transport. Sci. Adv., 6, eaax6869, https://doi.org/10.1126/sciadv.aax6869.

    • Search Google Scholar
    • Export Citation

  • Brodzik, M. J., D. G. Long, M. A. Hardman, A. Paget, and R. Armstrong, 2016: MEaSUREs calibrated enhanced-resolution passive microwave daily ease-grid 2.0 brightness temperature ESDR, version 1. NASA National Snow and Ice Data Center Distributed Active Archive Center, accessed 22 February 2020, https://doi.org/10.5067/measures/cryosphere/nsidc-0630.001.

  • Comiso, J. C., J. Yang, S. Honjo, and R. A. Krishfield, 2003: Detection of change in the Arctic using satellite and in situ data. J. Geophys. Res., 108, 3384, https://doi.org/10.1029/2002JC001347.

    • Search Google Scholar
    • Export Citation

  • Couture, G., 2022: Analyses spatiotemporelles des conditions de glace de mer et des tendances de formation des polynies de l’Archipel Arctique Canadien. M.S. thesis, Département de géomatique appliquée Faculté des lettres et sciences humaines, Université de Sherbrooke, 89 pp., https://savoirs.usherbrooke.ca/handle/11143/19074.

  • Cullather, R. I., Y.-K. Lim, L. N. Boisvert, L. Brucker, J. N. Lee, and S. M. Nowicki, 2016: Analysis of the warmest Arctic winter, 2015–2016. Geophys. Res. Lett., 43, 10 808–10 816, https://doi.org/10.1002/2016GL071228.

    • Search Google Scholar
    • Export Citation

  • Derksen, C., and R. Brown, 2012: Spring snow cover extent reductions in the 2008–2012 period exceeding climate model projections. Geophys. Res. Lett., 39, L19504, https://doi.org/10.1029/2012GL053387.

    • Search Google Scholar
    • Export Citation

  • Derksen, C., and L. Mudryk, 2023: Assessment of Arctic seasonal snow cover rates of change. Cryosphere, 17, 1431–1443, https://doi.org/10.5194/tc-17-1431-2023.

    • Search Google Scholar
    • Export Citation

  • Dolant, C., A. Langlois, B. Montpetit, L. Brucker, A. Roy, and A. Royer, 2016: Development of a rain-on-snow detection algorithm using passive microwave radiometry. Hydrol. Processes, 30, 3184–3196, https://doi.org/10.1002/hyp.10828.

    • Search Google Scholar
    • Export Citation

  • Dolant, C., A. Langlois, L. Brucker, A. Royer, A. Roy, and B. Montpetit, 2018a: Meteorological inventory of rain-on-snow events in the Canadian Arctic Archipelago and satellite detection assessment using passive microwave data. Phys. Geogr., 39, 428–444,https://doi.org/10.1080/02723646.2017.1400339.

    • Search Google Scholar
    • Export Citation

  • Dolant, C., B. Montpetit, A. Langlois, L. Brucker, O. Zolina, C. A. Johnson, A. Royer, and P. Smith, 2018b: Assessment of the barren ground Caribou die-off during winter 2015–2016 using passive microwave observations. Geophys. Res. Lett., 45, 4908–4916, https://doi.org/10.1029/2017GL076752.

    • Search Google Scholar
    • Export Citation

  • Dupont, F., G. Picard, A. Royer, M. Fily, A. Roy, A. Langlois, and N. Champollion, 2014: Modeling the microwave emission of bubbly ice: Applications to blue ice and superimposed ice in the Antarctic and Arctic. IEEE Trans. Geosci. Remote Sens., 52, 6639–6651, https://doi.org/10.1109/TGRS.2014.2299829.

    • Search Google Scholar
    • Export Citation

  • Foster, J. L., D. K. Hall, A. T. C. Chang, and A. Rango, 1984: An overview of passive microwave snow research and results. Rev. Geophys., 22, 195–208, https://doi.org/10.1029/RG022i002p00195.

    • Search Google Scholar
    • Export Citation

  • Gautier, C., 2022: Étude de la qualité d’habitat et des patrons migratoires du caribou de Peary (Rangifer tarandus pearyi) à l’aide d’un modèle thermodynamique de neige, des anomalies de glace de mer et du savoir traditionnel. M.S. thesis, Dépt. de géomatique appliquée Faculté des lettres et sciences humaines, Université de Sherbrooke, 155 pp., https://savoirs.usherbrooke.ca/handle/11143/18415.

  • Grenfell, T. C., and J. Putkonen, 2008: A method for the detection of the severe rain-on-snow event on Banks Island, October 2003, using passive microwave remote sensing. Water Resour. Res., 44, W03425, https://doi.org/10.1029/2007WR005929.

    • Search Google Scholar
    • Export Citation

  • Langlois, A., and Coauthors, 2017: Detection of rain-on-snow (ROS) events and ice layer formation using passive microwave radiometry: A context for Peary caribou habitat in the Canadian Arctic. Remote Sens. Environ., 189, 84–95, https://doi.org/10.1016/j.rse.2016.11.006.

    • Search Google Scholar
    • Export Citation

  • Larue, F., A. Royer, D. De Sève, A. Langlois, A. Roy, and L. Brucker, 2017: Validation of GlobSnow-2 snow water equivalent over eastern Canada. Remote Sens. Environ., 194, 264–277, https://doi.org/10.1016/j.rse.2017.03.027.

    • Search Google Scholar
    • Export Citation

  • Liston, G. E., and C. A. Hiemstra, 2011: The changing cryosphere: Pan-Arctic snow trends (1979–2009). J. Climate, 24, 5691–5712, https://doi.org/10.1175/JCLI-D-11-00081.1.

    • Search Google Scholar
    • Export Citation

  • Marmy, A., N. Salzmann, M. Scherler, and C. Hauck, 2013: Permafrost model sensitivity to seasonal climatic changes and extreme events in mountainous regions. Environ. Res. Lett., 8, 035048, https://doi.org/10.1088/1748-9326/8/3/035048.

    • Search Google Scholar
    • Export Citation

  • Martineau, C., 2020: Couplage du logiciel de modélisation de l’habitat MaxEnt à des simulations du couvert nival pour l’amélioration de la prédiction de présence du caribou de Peary. M.S. thesis, Département de géomatique appliquée Faculté des lettres et sciences humaines, Université de Sherbrooke, 92 pp., https://savoirs.usherbrooke.ca/handle/11143/17006.

  • Montpetit, B., A. Royer, A. Roy, A. Langlois, and C. Derksen, 2013: Snow microwave emission modeling of ice lenses within a snowpack using the microwave emission model for layered snowpacks. IEEE Trans. Geosci. Remote Sens., 51, 4705–4717, https://doi.org/10.1109/TGRS.2013.2250509.

    • Search Google Scholar
    • Export Citation

  • Moon, T. A., M. L. Druckenmiller, and R. L. Thoman, 2021: Arctic report card 2021: Executive summary. NOAA Tech. Rep. OAR ARC 21-01, 4 pp., https://doi.org/10.25923/5s0f-5163.

  • NOAA/ESRL/Physical Sciences Laboratory, 2020: NCEP North American regional reanalysis. NOAA, accessed 10 November 2020, https://psl.noaa.gov/data/gridded/data.narr.html.

  • Picard, G., L. Brucker, A. Roy, F. Dupont, M. Fily, A. Royer, and C. Harlow, 2013: Simulation of the microwave emission of multi-layered snowpacks using the Dense Media Radiative Transfer theory: The DMRT-ML model. Geosci. Model Dev., 6, 1061–1078, https://doi.org/10.5194/gmd-6-1061-2013.

    • Search Google Scholar
    • Export Citation

  • Putkonen, J., T. C. Grenfell, K. Rennert, C. Bitz, P. Jacobson, and D. Russell, 2009: Rain on snow: Little understood killer in the North. Eos, Trans. Amer. Geophys. Union, 90, 221–222, https://doi.org/10.1029/2009EO260002.

    • Search Google Scholar
    • Export Citation

  • Rantanen, M., A. Y. Karpechko, A. Lipponen, K. Nordling, O. Hyvärinen, K. Ruosteenoja, T. Vihma, and A. Laaksonen, 2022: The Arctic has warmed nearly four times faster than the globe since 1979. Commun. Earth Environ., 3, 168, https://doi.org/10.1038/s43247-022-00498-3.

    • Search Google Scholar
    • Export Citation

  • Rennert, K. J., G. Roe, J. Putkonen, and C. M. Bitz, 2009: Soil thermal and ecological impacts of rain on snow events in the circumpolar Arctic. J. Climate, 22, 2302–2315, https://doi.org/10.1175/2008JCLI2117.1.

    • Search Google Scholar
    • Export Citation

  • Riseborough, D., N. Shiklomanov, B. Etzelmüller, S. Gruber, and S. Marchenko, 2008: Recent advances in permafrost modelling. Permafrost Periglacial Processes, 19, 137–156, https://doi.org/10.1002/ppp.615.

    • Search Google Scholar
    • Export Citation

  • Roy, A., 2014: Modélisation de l’émission micro-onde hivernale en forêt boréale Canadienne. Ph.D. thesis, Université de Sherbrooke, 242 pp., http://savoirs.usherbrooke.ca/handle/11143/70.

  • Royer, A., F. Domine, A. Roy, A. Langlois, N. Marchand, and G. Davesne, 2021: New northern snowpack classification linked to vegetation cover on a latitudinal mega-transect across northeastern Canada. Écoscience, 28, 225–242, https://doi.org/10.1080/11956860.2021.1898775.

    • Search Google Scholar
    • Export Citation

  • Serreze, M. C., and R. G. Barry, 2011: Processes and impacts of Arctic amplification: A research synthesis. Global Planet. Change, 77, 85–96, https://doi.org/10.1016/j.gloplacha.2011.03.004.

    • Search Google Scholar
    • Export Citation

  • Shi, J., C. Xiong, and L. Jiang, 2016: Review of snow water equivalent microwave remote sensing. Sci. China Earth Sci., 59, 731–745, https://doi.org/10.1007/s11430-015-5225-0.

    • Search Google Scholar
    • Export Citation

  • Sokolov, A. A., N. A. Sokolova, R. A. Ims, L. Brucker, and D. Ehrich, 2016: Emergent rainy winter warm spells may promote boreal predator expansion into the Arctic. Arctic, 69, 121–129, https://doi.org/10.14430/arctic4559.

    • Search Google Scholar
    • Export Citation

  • Song, M., and J. Liu, 2017: The role of diminishing Arctic sea ice in increased winter snowfall over northern high-latitude continents in a warming environment. Acta Oceanol. Sin., 36, 34–41, https://doi.org/10.1007/s13131-017-1021-3.

    • Search Google Scholar
    • Export Citation

  • Weismüller, J., U. Wollschläger, J. Boike, X. Pan, Q. Yu, and K. Roth, 2011: Modeling the thermal dynamics of the active layer at two contrasting permafrost sites on Svalbard and on the Tibetan Plateau. Cryosphere, 5, 741–757, https://doi.org/10.5194/tc-5-741-2011.

    • Search Google Scholar
    • Export Citation
  • <sc>Fig</sc>. 1.
    View in gallery
    Fig. 1.

    Area covered by the seven island groups studied in this report.

  • <sc>Fig</sc>. 2.
    View in gallery
    Fig. 2.

    Overview of the algorithm validation results using data from meteorological stations. (a) Number of commissions in relation to the GRP threshold. (b) Number of omissions in relation to the GRP [Eq. (4)] threshold. (c) Error in percent [Eq. (5)] in relation to the GRP threshold.

  • <sc>Fig</sc>. 3.
    View in gallery
    Fig. 3.

    Annual average ROS occurrences from 1987 to 2019 with a fixed winter season from 1 Nov to 31 May as a scale of gray maxed at 4 ROS.

  • <sc>Fig</sc>. 4.
    View in gallery
    Fig. 4.

    Rain-on-snow occurrence anomalies map for winter 2010/11 using the fixed period method.

  • <sc>Fig</sc>. 5.
    View in gallery
    Fig. 5.

    Rain-on-snow occurrence anomalies map for winter 2015/16 using the fixed period method.

  • <sc>Fig</sc>. 6.
    View in gallery
    Fig. 6.

    Spatial average of ROS occurrence anomalies with the fixed period method. Blue represents a negative anomaly and orange represents a positive anomaly compared to the 1987–2018 average for each island group.

  • <sc>Fig</sc>. 7.
    View in gallery
    Fig. 7.

    Annual average ROS occurrences from 1987 to 2019 with the snow condition method as a scale of gray maxed at 4 ROS.

  • <sc>Fig</sc>. 8.
    View in gallery
    Fig. 8.

    Rain-on-snow occurrence anomalies map for winter 2010/11 using the snow condition method.

  • <sc>Fig</sc>. 9.
    View in gallery
    Fig. 9.

    Rain-on-snow occurrence anomalies map for winter 2015/16 using the snow condition method.

  • <sc>Fig</sc>. 10.
    View in gallery
    Fig. 10.

    Spatial average of ROS occurrence anomalies with the snow detection method. Blue refers to a negative anomaly and orange refers to a positive anomaly compared to the 1987–2018 average for each island group.

  • <sc>Fig</sc>. 11.
    View in gallery
    Fig. 11.

    Number of ROS events detected compared to the total area of open water around a buffer of 50 km from the island in October for each year. (a) Victoria Island. (b) Banks Island.

  • <sc>Fig</sc>. 12.
    View in gallery
    Fig. 12.

    Number of ROS events detected compared to the total area of open water around a buffer of 50 km from Victoria Island in October for each year: (a) northwest, (b) northeast, (c) southwest, and (d) southeast.

Fig. 1.

Area covered by the seven island groups studied in this report.
Fig. 2.

Overview of the algorithm validation results using data from meteorological stations. (a) Number of commissions in relation to the GRP threshold. (b) Number of omissions in relation to the GRP [Eq. (4)] threshold. (c) Error in percent [Eq. (5)] in relation to the GRP threshold.
Fig. 3.

Annual average ROS occurrences from 1987 to 2019 with a fixed winter season from 1 Nov to 31 May as a scale of gray maxed at 4 ROS.
Fig. 4.

Rain-on-snow occurrence anomalies map for winter 2010/11 using the fixed period method.
Fig. 5.

Rain-on-snow occurrence anomalies map for winter 2015/16 using the fixed period method.
Fig. 6.

Spatial average of ROS occurrence anomalies with the fixed period method. Blue represents a negative anomaly and orange represents a positive anomaly compared to the 1987–2018 average for each island group.
Fig. 7.

Annual average ROS occurrences from 1987 to 2019 with the snow condition method as a scale of gray maxed at 4 ROS.
Fig. 8.

Rain-on-snow occurrence anomalies map for winter 2010/11 using the snow condition method.
Fig. 9.

Rain-on-snow occurrence anomalies map for winter 2015/16 using the snow condition method.
Fig. 10.

Spatial average of ROS occurrence anomalies with the snow detection method. Blue refers to a negative anomaly and orange refers to a positive anomaly compared to the 1987–2018 average for each island group.
Fig. 11.

Number of ROS events detected compared to the total area of open water around a buffer of 50 km from the island in October for each year. (a) Victoria Island. (b) Banks Island.
Fig. 12.

Number of ROS events detected compared to the total area of open water around a buffer of 50 km from Victoria Island in October for each year: (a) northwest, (b) northeast, (c) southwest, and (d) southeast.
All Time Past Year Past 30 Days
Abstract Views 2062 1597 0
Full Text Views 735 555 338
PDF Downloads 225 94 42