• Armstrong, R., , and Brodzik M. , 1995: An earth-gridded SSM/I data set for cryospheric studies and global change monitoring. Adv. Space Res, 16 , 1015510163.

    • Search Google Scholar
    • Export Citation
  • Armstrong, R., , and Brodzik M. , 2001: Recent Northern Hemisphere snow extent: A comparison of data derived from visible and microwave satellite sensors. Geophys. Res. Lett, 28 , 36733676.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Armstrong, R., , and Brodzik M. , 2002: Northern Hemisphere EASE-Grid Weekly Snow Cover and Sea Ice Extent Version 2. National Snow and Ice Data Center, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Armstrong, R., , Knowles K. W. , , Brodzik M. J. , , and Hardman M. A. , 2003: DMSP SSM/I Pathfinder Daily EASE-Grid Brightness Temperatures. National Snow and Ice Data Center, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Briffa, K., , Jones P. , , and Schweingruber F. , 1992: Tree-ring density reconstructions of summer temperature patterns across western North America since 1600. J. Climate, 5 , 735754.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brown, R., 2000: Northern Hemisphere snow cover variability and change, 1915–97. J. Climate, 13 , 23392355.

  • Brown, R., , and Goodison B. , 1996: Interannual variability in reconstructed Canadian snow cover, 1915–1992. J. Climate, 9 , 12991318.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brown, R., , Brasnett B. , , and Robinson D. , 2003: Gridded North American monthly snow depth and snow water equivalent for GCM evaluation. Atmos.–Ocean, 41 , 114.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chang, A., , Foster J. , , and Hall D. , 1990: Satellite sensor estimates of Northern Hemisphere snow volume. Int. J. Remote Sens, 11 , 167171.

  • Derksen, C., , and Walker A. , 2003: Identification of systematic bias in the cross-platform (SMMR and SSM/I) EASE-Grid brightness temperature time series. IEEE Trans. Geosci. Remote Sens, 41 , 910915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Brown R. , , Walker A. , , and Brasnett B. , 2002a: Comparison of snow course, model and passive microwave derived snow water equivalent data for western North America. Proc. 59th Eastern Snow Conference, Stowe, VT, Eastern Snow Conference, 203– 213.

    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Walker A. , , LeDrew E. , , and Goodison B. , 2002b: Time series analysis of passive microwave derived central North American snow water equivalent imagery. Ann. Glaciol, 34 , 17.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Walker A. , , and Goodison B. , 2003a: A comparison of 18 winter seasons of in situ and passive microwave derived snow water equivalent estimates in western Canada. Remote Sens. Environ, 83 , 271282.

    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Walker A. , , LeDrew E. , , and Goodison B. , 2003b: Combining SMMR and SSM/I data for time series analysis of central North American snow water equivalent. J. Hydrometeor, 4 , 304316.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Foster, J., and Coauthors, 1996: Snow cover and snow mass intercomparisons of general circulation models and remotely sensed datasets. J. Climate, 9 , 409425.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Foster, J., , Chang A. , , and Hall D. , 1997: Comparison of snow mass estimates from a prototype passive microwave snow algorithm, a revised algorithm and a snow depth climatology. Remote Sens. Environ, 62 , 132142.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frei, A., , Miller J. , , and Robinson D. , 2003: Improved simulations of snow extent in the second phase of the Atmospheric Model Intercomparison Project (AMIP-2). J. Geophys. Res, 108 , 43694384.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goita, K., , Walker A. , , and Goodison B. , 2003: Algorithm development for the estimation of snow water equivalent in the boreal forest using passive microwave data. Int. J. Remote Sens, 24 , 10971102.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goodison, B., , and Walker A. , 1995: Canadian development and use of snow cover information from passive microwave satellite data. Passive Microwave Remote Sensing of Land–Atmosphere Interactions, B. Choudhury, et al., Eds., VSP BV, 245–262.

    • Search Google Scholar
    • Export Citation
  • Goodison, B., , Banga A. , , and Halliday R. , 1984: Canada–United States prairie snow cover runoff study. Can. Water Res. J, 9 , 99107.

  • Groisman, P., , Karl T. , , Knight R. , , and Stenchikov G. , 1994: Changes of snow cover, temperature, and radiative heat balance over the Northern Hemisphere. J. Climate, 7 , 16331656.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jezek, K., , Merry C. , , and Cavalieri D. , 1993: Comparison of SMMR and SSM/I passive microwave data collected over Antarctica. Ann. Glaciol, 17 , 131136.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kelly, R., , Chang A. , , Tsang L. , , and Foster J. , 2003: A prototype AMSR-E global snow area and snow depth algorithm. IEEE Trans. Geosci. Remote Sens, 41 , 230242.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Knowles, K., , Njoku E. , , Armstrong R. , , and Brodzik M. J. , 2002: Nimbus-7 SMMR Pathfinder Daily EASE-Grid Brightness Temperatures. National Snow and Ice Data Center, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Latif, M., , and Barnett T. , 1994: Causes of decadal variability over the North Pacific and North America. Science, 266 , 634637.

  • Leathers, D., , and Robinson D. , 1993: The association between extremes in North American snow cover extent and United States temperatures. J. Climate, 6 , 13451355.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Loveland, T., , Reed B. , , Brown J. , , Ohlen D. , , Zhu Z. , , Yang L. , , and Merchant J. , 2000: Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data. Int. J. Remote Sens, 21 , 13031330.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • MacKay, M., , and Derksen C. , 2003: Regional snowpack modeling over Canadian landscapes. Proc. 14th Int. Symp. and Workshop on Northern Research Basins, Kangerlussuaq/Sdr. Strømfjord, Greenland, 99–105.

    • Search Google Scholar
    • Export Citation
  • MacKay, M., , Seglenieks F. , , Verseghy D. , , Soulis E. , , Snelgrove K. , , Walker A. , , and Szeto K. , 2003: Modeling Mackenzie basin surface water balance during CAGES with the Canadian Regional Climate Model. J. Hydrometeor, 4 , 748767.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Meteorological Service of Canada, 2000: Canadian Snow Data. Climate Research Branch, CD-ROM.

  • Robinson, D., 1991: Merging operational satellite and historical station snow cover data to monitor climate change. Palaeogeogr., Palaeoclimatol., Palaeoecol, 90 , 235240.

    • Search Google Scholar
    • Export Citation
  • Robinson, D., , Dewey K. , , and Heim R. , 1993: Global snow cover monitoring: An update. Bull. Amer. Meteor. Soc, 74 , 16891696.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Robinson, D., , Tarpley J. , , and Ramsay B. , 1999: Transition from NOAA weekly to daily hemispheric snow charts. Preprints, 10th Symp. on Global Change Studies, Dallas, TX, Amer. Meteor. Soc., 487– 490.

    • Search Google Scholar
    • Export Citation
  • Serreze, M., and Coauthors, 2000: Observational evidence of recent change in the northern high-latitude environment. Climatic Change, 46 , 159207.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tait, A., 1998: Estimation of snow water equivalent using passive microwave radiation data. Remote Sens. Environ, 64 , 286291.

  • Walker, A., , and Goodison B. , 2000: Challenges in determining snow water equivalent over Canada using microwave radiometry. Proc. Int. Geoscience and Remote Sensing Symp., Honolulu, HI, WE09;ll03. Institute of Electrical and Electronics Engineers, CD-ROM, WE09_03.

    • Search Google Scholar
    • Export Citation
  • Walker, A., , and Silis A. , 2002: Snow cover variations over the Mackenzie River basin from SSM/I passive microwave satellite data. Ann. Glaciol, 34 , 814.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Walsh, J., , Tucek D. , , and Peterson M. , 1982: Seasonal snow cover and short-term climatic fluctuations over the United States. Mon. Wea. Rev, 110 , 14741485.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wiesnet, D., , Ropelewski C. , , Kukla G. , , and Robinson D. , 1987: A discussion of the accuracy of NOAA satellite-derived global seasonal snow cover measurements. Large Scale Effects of Seasonal Snow Cover, B. E. Goodison, R. G. Barry, and J. Dozier, Eds., Vol. 166, International Association of Hydrological Sciences, 291–304.

    • Search Google Scholar
    • Export Citation
  • Zhang, X., , Vincent L. , , Hogg W. , , and Nitsoo A. , 2000: Temperature and precipitation trends in Canada during the 20th century. Atmos.–Ocean, 38 , 395429.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • View in gallery

    Outline of the study domain

  • View in gallery

    (a) Average study area SWE retrievals, and relative frequency plots for the SWEunadjusted and SWEDW2003 datasets for (b) 1 Dec, (c) 1 Jan, and (d) 1 Feb 1978–87

  • View in gallery

    (a) Snow course measurement locations. (b) Frequency of passive microwave–derived SWE over- and underestimation relative to snow course measurements

  • View in gallery

    (a) Absolute difference between snow course SWE measurements and passive microwave SWE retrievals. (b) Average difference. For snow course measurement locations, refer to Fig. 3a

  • View in gallery

    Standardized anomaly sequence for the (a) SWEunadjusted and (b) SWEDW03 time series. Dashed line denotes transition from SMMR to SSM/I data. See Fig. 1 for study area domain

  • View in gallery

    (a) General classification agreement between NOAA snow extent charts and passive microwave–derived snow extent, 1978–2001. (b) Errors attributed to passive microwave omission and (c) commission

  • View in gallery

    Interannual variability in monthly SCE anomalies, standardized to a 1978–92 reference period. See Fig. 1 for study area domain

  • View in gallery

    Interannual variability in monthly SWE anomalies, standardized to a 1978–92 reference period. See Fig. 1 for study area domain

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 78 78 5
PDF Downloads 7 7 0

Merging Conventional (1915–92) and Passive Microwave (1978–2002) Estimates of Snow Extent and Water Equivalent over Central North America

View More View Less
  • 1 Climate Research Branch, Meteorological Service of Canada, Downsview, Ontario, Canada
  • | 2 Canadian Meteorological Centre, Montreal, Quebec, Canada
  • | 3 Climate Research Branch, Meteorological Service of Canada, Downsview, Ontario, Canada
© Get Permissions
Full access

Abstract

A detailed evaluation of snow water equivalent (SWE) and snow cover extent (SCE) derived using the combined Scanning Multichannel Microwave Radiometer (SMMR) and Special Sensor Microwave Imager (SSM/I) brightness temperature records for the 1978–2002 period was carried out for a longitudinal transect in the continental interior of North America. Comparison with in situ SWE observations showed that the SMMR brightness temperature adjustments are required to produce SWE retrievals with similar bias and rmse as observed during the SSM/I period. Underestimation of SCE in the passive microwave dataset (relative to NOAA snow charts) was identified as a systematic problem, most pronounced in early winter and during seasons with above-average snow extent. The passive microwave data were successfully merged with historical data based on strong interdataset agreement for a 1978–92 overlap period. Analysis of SWE and SCE time series for the months of December through March 1915–2002 provided information on SWE and SCE variability over the interior of North America, but yielded no evidence of significant trends in either variable over the 88-yr period. Anomalies observed during the relatively recent period of passive microwave data acquisition did not exceed the range of anomalies observed in the historical data record.

Corresponding author address: C. Derksen, Climate Research Branch, Meteorological Service of Canada, 4905 Dufferin Street, Downsview, ON M3H 5T4, Canada. Email: Chris.Derksen@ec.gc.ca

Abstract

A detailed evaluation of snow water equivalent (SWE) and snow cover extent (SCE) derived using the combined Scanning Multichannel Microwave Radiometer (SMMR) and Special Sensor Microwave Imager (SSM/I) brightness temperature records for the 1978–2002 period was carried out for a longitudinal transect in the continental interior of North America. Comparison with in situ SWE observations showed that the SMMR brightness temperature adjustments are required to produce SWE retrievals with similar bias and rmse as observed during the SSM/I period. Underestimation of SCE in the passive microwave dataset (relative to NOAA snow charts) was identified as a systematic problem, most pronounced in early winter and during seasons with above-average snow extent. The passive microwave data were successfully merged with historical data based on strong interdataset agreement for a 1978–92 overlap period. Analysis of SWE and SCE time series for the months of December through March 1915–2002 provided information on SWE and SCE variability over the interior of North America, but yielded no evidence of significant trends in either variable over the 88-yr period. Anomalies observed during the relatively recent period of passive microwave data acquisition did not exceed the range of anomalies observed in the historical data record.

Corresponding author address: C. Derksen, Climate Research Branch, Meteorological Service of Canada, 4905 Dufferin Street, Downsview, ON M3H 5T4, Canada. Email: Chris.Derksen@ec.gc.ca

1. Introduction

Synoptic and seasonal-scale variations in snow cover extent, depth, and water equivalent have important impacts on atmospheric energy and moisture exchanges, and on the terrestrial water balance. Spaceborne passive microwave data are well suited to regional monitoring of snow cover because of characteristics such as all-weather imaging, a wide swath width with frequent overpass times, and a long available time series. Perhaps most significantly, the scattering influence of dry snow cover allows the estimation of snow water equivalent (SWE), a variable that cannot be derived from optical spaceborne imagery. Over the past 10 yr, the Climate Research Branch of the Meteorological Service of Canada (MSC) has developed a suite of land-cover-sensitive passive microwave SWE algorithms for western Canada (Goodison and Walker 1995; Goita et al. 2003) that are used to generate weekly SWE maps over the Canadian prairie and boreal forest regions for a number of operational users including hydroelectric companies, water resource management agencies, and weather forecast offices. The regional, land-cover-sensitive approach adopted by the MSC (see Walker and Goodison 2000) differs from the global-scale perspective of most other contemporary passive microwave SWE algorithms (i.e., Tait 1998; Kelly et al. 2003; Chang et al. 1990; Foster et al. 1997). A current limitation of this regional approach is a constrained study area limited to the open prairies and boreal forest of western Canada, and the Great Plains of the United States—land cover types for which specific field campaigns have supported algorithm development and evaluation (Goita et al. 2003; Goodison et al. 1984). A strong benefit, however, is a focused, relatively data rich area for evaluation, applications, and feedback from operational users (see Derksen et al. 2002b; 2003a,b).

The period of available multifrequency spaceborne passive microwave brightness temperatures includes Scanning Multichannel Microwave Radiometer (SMMR) data from 1978 to 1987 and Special Sensor Microwave Imager (SSM/I) data from 1987 to present. The SMMR and SSM/I data are available in a common gridded format [the Equal Area Scalable Earth Grid (EASE-Grid); see Armstrong and Brodzik 1995] from the National Snow and Ice Data Center (NSIDC; Knowles et al. 2002; Armstrong et al. 2003). The sensors have slightly different spatial, temporal, and radiometric characteristics (Table 1) that impact the continuity and consistency of cross-platform brightness temperatures and derived geophysical variables. When unadjusted EASE-Grid brightness temperatures are utilized, SWE and snow cover extent (SCE) retrievals during SMMR seasons are systematically and significantly lower than retrievals during SSM/I seasons (Derksen et al. 2003b). Jezek et al. (1993) developed cross-platform brightness temperature standardization coefficients using daily averaged data from Antarctica; however, these adjustments are not necessarily appropriate for midlatitude terrestrial applications with diurnal and hence overpass-time-sensitive properties such as snow cover.

Derksen and Walker (2003) derived new regression-based coefficients with sensitivity to overpass time using collocated brightness temperatures over terrestrial surfaces of central North America from the August 1987 period of overlap between the SMMR and the first SSM/I onboard the U.S. Defense Meteorological Satellite Program (DMSP) F-8 satellite. This procedure adjusted SMMR brightness temperatures to an SSM/I F-8 baseline, reducing the mean cross-platform brightness temperature offset from nearly 9 K to approximately 0.5 K for the 18- and 19-V channels, and from 4.5 to 0.4 K for the two 37-V channels. This improvement in agreement was the result of increasing the SMMR 18- and 37-V cold overpass brightness temperatures by an average of 7.5 and 2.0 K, respectively, although the absolute adjustments are dependant on brightness temperature magnitude (see Derksen and Walker 2003).

The purpose of the present study was to evaluate the brightness temperature homogenization approach of Derksen and Walker (2003) and to develop consistent cross-platform SCE and SWE time series from passive microwave data over the continental interior of North America (Fig. 1). This particular region has been used in a number of previous studies to develop reliable passive microwave–derived SWE estimates (Derksen et al. 2002a,b, 2003a,b). The study area provides a north– south transect through three important land cover classes (open prairie, mixed transitional, and boreal forest), and the authors have a high level of confidence in the retrievals for this domain. Climatologically, the study area includes the dominant center of North American snow cover variability as identified by Groisman et al. (1994). Prairie snow cover anomalies are strongly correlated with surface temperature departures (Walsh et al. 1982), while Leathers and Robinson (1993) identify Great Plains snow cover, or the lack of it, as a significant modifier of winter air masses moving south out of Canada. Finally, the surface observing network is relatively dense within this region (especially south of 55°N latitude), providing useful information for passive microwave algorithm assessment and the second objective of this study: combining the passive microwave and historical conventional snow cover data records.

The merging of satellite and historical conventional information on snow cover and SWE is required to provide consistent long-term time series for documenting and understanding climate variability and change (see Serreze et al. 2000), as well as for evaluating climate models (see Foster et al. 1996; Frei et al. 2003). This need is further underscored by worldwide trends toward reductions in surface observing networks. For example, at peak levels in the early 1980s there were over 1700 snow courses operating in Canada. This number declined to around 800 in the early 1990s (Brown 2000). The satellite-derived data (1978–2002) were therefore combined with gridded historical (1915–92) estimates of SCE and SWE for midlatitudinal North America (Brown 2000) to provide an 88-yr record for examining trends and variability. An extensive comparison and validation was carried out of the 1978–92 period of overlap between the two data series.

2. Data

a. Passive microwave–derived SWE data

The Climate Research Branch of the MSC has an ongoing program to develop SWE retrieval capabilities from spaceborne passive microwave brightness temperatures for major Canadian landscape regions (Walker and Goodison 2000). Initial MSC algorithm development focused on the open, generally nonforested environment of the Canadian prairies (Goodison and Walker 1995), while subsequent research produced algorithms for general categories of boreal forest cover (Goita et al. 2003). The open prairie algorithm is based on the brightness temperature gradient between the 37- and 19-GHz (18 GHz with SMMR) vertically polarized channels, while the three forest algorithms are based on the brightness temperature difference of these same channels. The resulting suite of land-cover-sensitive SWE retrieval algorithms can be applied to both SMMR and SSM/I brightness temperatures, with per-grid-cell SWE estimates produced as the sum of the SWE values obtained from each land cover algorithm weighted by the percentage land cover type (F) within each grid cell:
DDCCSSOO
where D represents deciduous forest, C coniferous forest, S sparse forest, and O open prairie environments. Grid cell land cover fractions are determined from the International Geosphere–Biosphere Programme (IGBP) 1-km global land cover classification (Loveland et al. 2000), resampled to the EASE-Grid by NSIDC. Full details on algorithm development and initial validation are provided in Goodison and Walker (1995) and Goita et al. (2003). Evaluation of the MSC algorithm suite for various regions and time series is described in Derksen et al. (2002b; 2003a,b) and Walker and Silis (2002). The algorithms are typically capable of producing SWE retrievals within ±15 mm of surface observations, although consistent underestimation of SWE is a problem in heavily forested areas because of the complex impact of dense vegetation on microwave emission and scatter (Walker and Silis 2002; Derksen et al. 2002b, 2003a).

All the brightness temperature frequencies have grid cell dimensions of 25 km × 25 km, although the microwave emission is measured from a larger, elliptical field of view and resampled [see Armstrong and Brodzik (1995) for a complete description of the EASE-Grid resampling procedure]. Two orbits of data are acquired daily (ascending and descending overpasses), although the SMMR was deactivated every other day as a power saving measure. For the present study, pentad resolution (5-day averaged), 24-season (1 December to 1 March, 1978–2002) SCE and SWE time series were derived. Before processing with the MSC algorithm suite, SMMR brightness temperatures were standardized to an SSM/I F-8 baseline using the frequency, polarization, and overpass-time-specific coefficients of Derksen and Walker (2003). A second passive microwave–derived SWE dataset was also produced for comparison, applying no adjustments to the SMMR brightness temperatures. These two datasets will subsequently be referred to as SWEDW2003 and SWEunadjusted, respectively. For both datasets, SMMR and SSM/I brightness temperatures from cold overpass times (midnight for SMMR; 0600 LT for SSM/I) were utilized to increase the frequency of monitoring a cold and dry snowpack—conditions that optimize algorithm performance. SCE was determined by classifying each grid cell with SWE greater than 1 mm as snow covered. This threshold is consistent with the findings of Robinson (1991) and Brown (2000) that in open terrain, the NOAA visible snow cover product corresponds to shallow snow conditions of 1–2 cm in depth.

b. Snow course data

Snow course (survey) data are collected by a number of national and provincial agencies across Canada, and a CD-ROM compilation of data up to 1995 has been generated by the Climate Research Branch (Meteorological Service of Canada 2000). Many of the observing programs only collect data during the second half of the water year to obtain an estimate of maximum SWE prior to melt. MSC snow surveys, however, are carried out on a regular (weekly/biweekly) basis throughout the snow cover season. This study used a distributed network of 15 MSC snow courses for comparison with the passive microwave datasets.

c. Historical SWE and SCE estimates

Historical gridded estimates of SCE and SWE over the midlatitudes (∼40° to 60°N) of North America were derived by Brown (2000) from in situ snow depth observations from Canada and the United States over the 1915–92 period. SWE estimates were derived by assuming a fixed within-season variation in snow density based on an analysis of snow course data from 1964 to 1993. Brown (2000) demonstrated that interannual variability in snow density was low over the midlatitudes, which justified the application of mean 2-week-averaged density values to derive SWE from depth measurements, with unique mean densities derived for prairie and boreal forest environments. An inverse-distance interpolation method was used to grid the SCE and SWE estimates to a 190.5-km-resolution polar stereographic projection. The spatial interpolation did not take topographic variability into account, but this is not a major concern over the region examined in this study where topographic variation is limited.

d. Weekly NOAA snow extent charts

Weekly snow extent charts have been available since 1966 with the launch of the first National Oceanic and Atmospheric Administration (NOAA) meteorological satellite (see Wiesnet et al. 1987). A detailed description of the digitized version of this dataset is provided by Robinson et al. (1993). This study used a recent release of the NOAA weekly snow chart dataset interpolated to EASE-Grid format by Armstrong and Brodzik (2002) to facilitate the comparison with passive microwave– derived snow cover parameters. Extensive work has been carried out by Robinson to ensure the NOAA dataset is homogeneous up to June 1999 when the charting process was replaced by the higher-resolution (25 km) daily Interactive Multisensor Snowmap (IMS) product (Robinson et al. 1999). The IMS product uses a blend of visible imagery, passive microwave information, and in situ observations.

3. Results

a. Homogenization of passive microwave–derived SWE data

The equations derived by Derksen and Walker (2003) for standardizing cold overpass EASE-Grid SMMR 18- and 37-V brightness temperatures to SSM/I F-8 19- and 37-V data are shown in Eqs. (2) and (3):
i1525-7541-5-5-850-e2
Of primary interest is how these adjustments account and correct for the bias in SWE retrievals during the SMMR period. A comparison with the SWEunadjusted dataset shows that the SWEDW2003 retrievals are higher across the study area (Fig. 2a), indicating that the adjusted brightness temperature dataset corrects the systematic SMMR SWE underestimation described in Derksen et al. (2003b). On 1 December, the SWEDW2003 average is approximately 7 mm higher than the SWEunadjusted average, and by 1 February the difference reaches 15 mm. Relative frequency plots for three dates through the winter season (Figs. 2b–d) show two important differences between the SWEunadjusted and SWEDW2003 datasets: a shift in the SWEDW2003 distribution toward higher SWE retrievals, and a notable increase in the number of snow-covered grid cells (SWE retrievals >0). A comparative analysis with independent SWE and SCE datasets was subsequently undertaken to establish the relative accuracies of these two datasets.

b. Evaluation with in situ SWE estimates

Time series from the two passive microwave datasets (SWEunadjusted and SWEDW2003) were compared to a spatially distributed network of snow course SWE measurements through the SMMR time period (1978/79 through 1986/87) to further assess the impact of the Derksen and Walker (2003) brightness temperature adjustments on SWE retrievals. A point-to-point comparison was made for each EASE-Grid cell within which an MSC snow course measurement site was located. There are some obvious scale issues with an evaluation of this nature given the large dimensions of the gridded passive microwave data; however, snow course sites were established to characterize the surrounding land cover, and the long time series available for comparison provides some insight on general dataset agreement and any potential systematic bias. The snow course measurements were compared to the SWE retrievals for three consistent measurement dates through the time series (1 December, 1 January, and 1 February). The frequency of passive microwave–derived SWE overestimation versus underestimation relative to the snow course data is shown in Fig. 3. The problem with using unadjusted EASE-Grid SMMR brightness temperatures is readily apparent: approximately 80% of SMMR SWEunadjusted retrievals are lower than the surface measurements. Conversely, unadjusted SSM/I brightness temperatures exhibit no strong bias between over- and underestimation (Fig. 3). The tendency for SMMR SWE retrievals to be too low is removed in the SWEDW2003 dataset, with the bias statistics very similar to that observed within the SSM/I data record. Absolute and average bias values for the passive microwave retrievals and snow course measurements are shown in Fig. 4. The SWEDW2003 dataset exhibits an improvement of approximately 5 mm in the absolute bias, and the average bias values near zero illustrate that the issue of consistent SWE underestimation has been largely removed from the dataset. It should be noted, however, that the majority of snow course locations (Fig. 3) are located north of the marginal snow cover zone in the southern prairies, where microwave SCE underestimation is most pronounced. There is a tendency toward SWE underestimation in December (Fig. 4), which does not persist as the winter season progresses into January and February, showing that SWE retrieval errors do not compound in a cumulative fashion.

These results indicate that the brightness temperature adjustments of Derksen and Walker (2003) produce SWE retrievals during the SMMR time period with similar accuracy and bias characteristics to those observed during the SSM/I time series, creating a homogeneous cross-platform time series. A SWE anomaly time series for the SWEunadjusted and SWEDW2003 datasets was computed (Fig. 5), using pentad statistics for winter seasons (December, January, and February) spanning 1978– 2002. In the unadjusted time series, the SMMR record is dominated by negative SWE anomalies, while the opposite is true for the SSM/I period. In the SWEDW2003 time series, no systematic anomaly bias is evident across the two platforms. The anomaly values for the seasons of 1997/98 through 1999/2000 warrant special attention. In the unadjusted time series, the anomalies indicate near-normal SWE values. These seasons, however, were characterized by below-normal winter precipitation and deficit snow depth over much of the study domain. The correspondingly low SWE values across the domain are realistically reflected as negative SWE anomalies after the brightness temperature adjustments of Derksen and Walker (2003) are applied.

c. Evaluation with NOAA snow charts

An evaluation of SCE estimates derived from the SWEDW2003 dataset was carried out using NOAA weekly binary snow extent charts in the EASE-Grid (Armstrong and Brodzik 2002). Twenty-three winter seasons were included in the evaluation: 1978/79 through 2000/01. This includes the post–June 1999 period with the IMS product, but the impact of any inhomogeneity is not expected to be large over the study domain as the snow line is well defined given the study area and season of focus. For this comparison, EASE-Grid cells with passive microwave–derived SWE greater than or equal to 1 mm were assumed to be snow covered, while SWE values less than 1 mm were considered snow free.

Classification agreement patterns (percentage of dates with the same snow extent classification, Fig. 6a) show good agreement with the exception of the marginal snow cover zone, where snow cover can be patchy, thin, and potentially wet. Under such conditions, the passive microwave SWE algorithm will produce zero or negligible SWE retrievals. General agreement is very high across the forested regions of northern Saskatchewan and Manitoba, Canada, with the proportion of the study area with nearly 100% agreement between datasets increasing from 1 December–1 February as the snow line advances southward. The impact of land cover on SWE algorithm performance is apparent in the general classification results. The coniferous algorithm applied to the forested regions in the north of the study area retrieves SWE greater than 1 mm for all time periods in the time series, producing SCE patterns that agree strongly with the NOAA product. Immediately south of the boreal forest, the open prairie algorithm does not always identify snow cover in the early winter: this disparity causes the tree line to appear as a feature in the 1 December results in Fig. 6a. The direction of this disagreement between datasets in the open prairie is illustrated clearly in the omission error patterns in Fig. 6b. Snow extent omission errors in the passive microwave dataset (NOAA = snow; passive microwave = no snow) are common in early winter because the snowpack is characteristically shallow and of low density, so volume scatter of the microwave radiation is negligible. This finding is consistent with the tendency for SWE retrievals to be too low in December, even after the SMMR brightness temperature adjustments are applied (see Fig. 4b). By the end of the winter season, snow extent omission and commission errors are relatively mixed (Figs. 6b,c).

Passive microwave and NOAA dataset agreement, expressed as the percentage of grid cells in both datasets with the same snow cover classification, is significantly correlated with the occurrence of passive microwave omission errors for all four time intervals examined, while correlation with commission errors is low and insignificant (Table 2). This confirms the tendency for passive microwave–derived datasets to underestimate snow extent when compared with optically derived products. This systematic bias was also identified through the NOAA snow chart data record by Armstrong and Brodzik (2001), and in the relatively short Moderate Resolution Imaging Spectroradiometer (MODIS; data available since 1999) data record by Bussieres et al. (2002). There is some predictability to dataset disagreement: classification errors are positively and significantly correlated with passive microwave snow extent anomalies during December and January (Table 2).

Comparisons between passive microwave SWE retrievals and in situ snow course measurements, and passive microwave–derived snow extent and NOAA snow charts were performed to assess the accuracy of a cross-platform (SMMR and SSM/I) time series. The SMMR brightness temperature adjustments of Derksen and Walker (2003) remove the systematic underestimation of SWE when unadjusted SMMR brightness temperatures are utilized. When the SWEDW2003 dataset is converted to snow extent, winter-season classification agreement with NOAA snow charts ranges spatially from approximately 65% in the marginal snow cover zone to 100% in the northern boreal forest. Therefore, the cross-platform SWEDW2003 dataset is considered suitable for addressing research issues that demand a rigorous time series, and the next section of this paper provides information on SWE variability and change in western North America.

d. Merging of satellite and historical SCE and SWE data

To combine the historical and satellite SCE and SWE time series, monthly averages were constructed over the study area and converted to standardized anomalies with respect to the 15-yr period of overlap (1978–92) between the two datasets. A summary of the monthly SCE and SWE correlations between the two data series is given in Table 3. Both variables were significantly correlated over the December to March period, with the highest SCE correlation in December (r = 0.93), highest SWE correlation in February (r = 0.91), and moderately strong correlations for the complete winter season (SCE: r = 0.85; SWE: r = 0.81). The coefficient of variation (CV; standard deviation/mean) increases from December through February in both time series, although greater interannual variability is characterized by the conventional dataset (Table 4). The CV values are insignificantly different at the 95% level. Coefficient of variation statistics produced from an independent comparison with a SWE reanalysis dataset (Brown et al. 2003) over the same study domain and time period confirmed slightly reduced interannual variability in the passive microwave–derived dataset (Derksen et al. 2002a).

The historical and passive microwave–derived anomaly time series for December through March are shown in Figs. 7 (SCE) and 8 (SWE). It is very important to note that anomalies from the spaceborne passive microwave era clearly fall within the anomaly range observed in the historical data record back to 1915. A noticeable feature of the monthly SWE series for January and February is an approximately 20-yr cycle, which was also identified in winter snow cover over western Canada by Brown and Goodison (1996), who related this source of variability to the 20-yr subtropical gyre circulation of the North Pacific (Latif and Barnett 1994). Linear trend analysis provided no evidence of any significant long-term trends in either SCE or SWE (Table 5). Serial correlation through each monthly anomaly time series is low (<0.1) so no adjustments were made before trend analysis was performed. Positive trends in the combined time series are a function of the largely deficit SCE and SWE conditions that dominate the first 30 yr of the time series. Positive trends of a reduced magnitude were found when the conventional and passive microwave datasets were considered separately. The only negative trends were found within the passive microwave SCE dataset, with the trend in this time series strongly influenced by strong positive departures during the first winter (1978/79) of the satellite time series.

4. Discussion and conclusions

The Meteorological Service of Canada has adopted a regional approach to deriving SWE retrieval algorithms for passive microwave brightness temperatures. Land cover and snowpack controls on algorithm performance result in spatial (open prairies and boreal forest) and temporal (winter season) constraints on algorithm performance. With the long record of spaceborne passive microwave data, however, there is great potential in this dataset for various hydrological, climatological, and numerical modeling applications.

Brightness temperature adjustments are necessary to produce homogeneous cross-platform (SMMR and SSM/I) passive microwave–derived SCE and SWE time series (Derksen et al. 2003a,b). The first objective of this study was to evaluate the SMMR brightness temperature adjustments of Derksen and Walker (2003) with respect to their impact on winter-season SWE retrievals and SCE estimates made with the MSC algorithm suite. Evaluation with snow course measurements shows persistent SWE underestimation during SMMR seasons is corrected using the adjustments of Derksen and Walker (2003), producing a time series with evenly distributed bias statistics (over- versus underestimation of SWE relative to snow course measurements).

Three general statements can be made based on the results of SCE classification comparison between the passive microwave time series [using the Derksen and Walker (2003) adjustments] and NOAA snow charts. First, errors in passive microwave snow extent (and by extension, SWE) mapping are caused by an underestimation of SCE, which is most pronounced in early winter. Disagreement between the two datasets is strongly driven by the passive microwave dataset classifying as snow free grid cells that are snow covered in the NOAA dataset. Spatially, the region of maximum disagreement between the two datasets follows the southward advance of the marginal snow cover zone through the winter. Second, the passive microwave and NOAA datasets agree very strongly for this study area for 1 February, a function of both datasets identifying extensive snow extent for this time period. SCE classification errors in the boreal forest are not a problem in the winter, but likely would be evident if this analysis was extended to the shoulder seasons of spring and fall. Third, interannual variability in dataset agreement can be linked to early-winter SCE, with passive microwave snow extent underestimation greater during seasons with positive SCE anomalies in the NOAA data record. Early-winter SCE can therefore be used as a predictor of passive microwave SCE estimation accuracy: December and January passive microwave omission errors are a more significant problem during seasons with anomalously extensive SCE.

Passive microwave–derived SCE and SWE anomalies agree well with anomalies from historical gridded datasets over a 1978–92 overlap period. This agreement provides confirmation of passive microwave algorithm performance based on the strong correlations derived during the overlap period and allows the passive microwave data record to be viewed within the context of a much longer time series. The passive microwave–derived SCE and SWE anomalies over the past quarter century fall within the range of observed anomalies calculated back to 1915 from the historical gridded SCE and SWE datasets described in Brown (2000). There was no evidence of any significant linear trends in SWE or SCE over the study region for the 88-yr period, although the SWE series did display evidence of a 20-yr cycle that is a common feature of climatological time series in western North America (Briffa et al. 1992).

These trend results can also be placed in the context of gridded temperature and precipitation dataset analysis for southern Canada (<60°N) by Zhang et al. (2000) who found a warming trend of 0.5° to 1.5°C between 1900 and 1998 across the area used in this study. This trend was driven by a decreased diurnal temperature range caused by warming daily minimum temperatures. Zhang et al. (2000) also isolated an increased ratio of snowfall to annual total precipitation. The results of the present study indicate that there is little net impact of the temperature and precipitation trends identified by Zhang et al. (2000) on SWE during the winter season.

Positive SWE anomalies can be caused by an anomalously extensive snow-covered area, or an increase in SWE over the climatologically normal snow-covered area. Likewise, deficit SWE anomalies can be caused by a retreat in snow extent, or a thinning of the water equivalent over the normal snow-covered area. An investigation of coincident SCE and SWE anomalies may prove to be an interesting starting point for examining atmospheric drivers behind snow cover anomalies in future studies. It is reasonable to hypothesize that different atmospheric mechanisms (synoptic circulation; storm tracks) are behind the strongly positive SWE anomaly of March 1969 when the SCE anomaly is approximately +2, and the strongly positive SWE anomaly of March 1973 when the SCE anomaly is approximately −1. Lagged relationships between early-season SCE and later-season SWE could also have some utility as a water resource forecasting tool.

Together, the SWE and SCE evaluations presented in this study illustrate that the MSC land-cover-sensitive passive microwave SWE algorithm suite yields a time series of data that can be merged with historical datasets. The MSC algorithms are used to produce near-real-time SWE maps for flow and flood forecasting, drought assessment, and spring soil moisture recharge potential. The passive microwave time series is also contributing to a number of ongoing applications at MSC, including the identification of synoptic-scale relationships between SWE distribution and atmospheric circulation, and evaluation of snow cover processes and modeled SWE fields in the Canadian Regional Climate Model (MacKay et al. 2003; MacKay and Derksen 2003). Recent research on algorithm performance in the mixed forest environments of the Mackenzie basin (MacKay et al. 2003) and broadleaf forest and tundra regions (MacKay and Derksen 2003) will allow passive microwave SWE methods to be applied to investigate SCE and SWE variability over a larger geographic area. While the region examined in this study was appropriate for evaluation of the brightness temperature standardization, and the first attempt at merging passive microwave and conventional time series, a deeper climatological perspective will be gained by performing trend and time series analysis across a more extensive domain.

Acknowledgments

The EASE-Grid data (brightness temperatures, NOAA snow extent data, land cover information) were obtained from the EOSDIS National Snow and Ice Data Center Distributed Active Archive Center (NSIDC DAAC), University of Colorado, Boulder. This research is a contribution to the Canadian CRYSYS project. Thanks to three anonymous reviewers for their constructive comments on this manuscript.

REFERENCES

  • Armstrong, R., , and Brodzik M. , 1995: An earth-gridded SSM/I data set for cryospheric studies and global change monitoring. Adv. Space Res, 16 , 1015510163.

    • Search Google Scholar
    • Export Citation
  • Armstrong, R., , and Brodzik M. , 2001: Recent Northern Hemisphere snow extent: A comparison of data derived from visible and microwave satellite sensors. Geophys. Res. Lett, 28 , 36733676.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Armstrong, R., , and Brodzik M. , 2002: Northern Hemisphere EASE-Grid Weekly Snow Cover and Sea Ice Extent Version 2. National Snow and Ice Data Center, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Armstrong, R., , Knowles K. W. , , Brodzik M. J. , , and Hardman M. A. , 2003: DMSP SSM/I Pathfinder Daily EASE-Grid Brightness Temperatures. National Snow and Ice Data Center, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Briffa, K., , Jones P. , , and Schweingruber F. , 1992: Tree-ring density reconstructions of summer temperature patterns across western North America since 1600. J. Climate, 5 , 735754.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brown, R., 2000: Northern Hemisphere snow cover variability and change, 1915–97. J. Climate, 13 , 23392355.

  • Brown, R., , and Goodison B. , 1996: Interannual variability in reconstructed Canadian snow cover, 1915–1992. J. Climate, 9 , 12991318.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brown, R., , Brasnett B. , , and Robinson D. , 2003: Gridded North American monthly snow depth and snow water equivalent for GCM evaluation. Atmos.–Ocean, 41 , 114.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chang, A., , Foster J. , , and Hall D. , 1990: Satellite sensor estimates of Northern Hemisphere snow volume. Int. J. Remote Sens, 11 , 167171.

  • Derksen, C., , and Walker A. , 2003: Identification of systematic bias in the cross-platform (SMMR and SSM/I) EASE-Grid brightness temperature time series. IEEE Trans. Geosci. Remote Sens, 41 , 910915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Brown R. , , Walker A. , , and Brasnett B. , 2002a: Comparison of snow course, model and passive microwave derived snow water equivalent data for western North America. Proc. 59th Eastern Snow Conference, Stowe, VT, Eastern Snow Conference, 203– 213.

    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Walker A. , , LeDrew E. , , and Goodison B. , 2002b: Time series analysis of passive microwave derived central North American snow water equivalent imagery. Ann. Glaciol, 34 , 17.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Walker A. , , and Goodison B. , 2003a: A comparison of 18 winter seasons of in situ and passive microwave derived snow water equivalent estimates in western Canada. Remote Sens. Environ, 83 , 271282.

    • Search Google Scholar
    • Export Citation
  • Derksen, C., , Walker A. , , LeDrew E. , , and Goodison B. , 2003b: Combining SMMR and SSM/I data for time series analysis of central North American snow water equivalent. J. Hydrometeor, 4 , 304316.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Foster, J., and Coauthors, 1996: Snow cover and snow mass intercomparisons of general circulation models and remotely sensed datasets. J. Climate, 9 , 409425.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Foster, J., , Chang A. , , and Hall D. , 1997: Comparison of snow mass estimates from a prototype passive microwave snow algorithm, a revised algorithm and a snow depth climatology. Remote Sens. Environ, 62 , 132142.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Frei, A., , Miller J. , , and Robinson D. , 2003: Improved simulations of snow extent in the second phase of the Atmospheric Model Intercomparison Project (AMIP-2). J. Geophys. Res, 108 , 43694384.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goita, K., , Walker A. , , and Goodison B. , 2003: Algorithm development for the estimation of snow water equivalent in the boreal forest using passive microwave data. Int. J. Remote Sens, 24 , 10971102.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goodison, B., , and Walker A. , 1995: Canadian development and use of snow cover information from passive microwave satellite data. Passive Microwave Remote Sensing of Land–Atmosphere Interactions, B. Choudhury, et al., Eds., VSP BV, 245–262.

    • Search Google Scholar
    • Export Citation
  • Goodison, B., , Banga A. , , and Halliday R. , 1984: Canada–United States prairie snow cover runoff study. Can. Water Res. J, 9 , 99107.

  • Groisman, P., , Karl T. , , Knight R. , , and Stenchikov G. , 1994: Changes of snow cover, temperature, and radiative heat balance over the Northern Hemisphere. J. Climate, 7 , 16331656.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jezek, K., , Merry C. , , and Cavalieri D. , 1993: Comparison of SMMR and SSM/I passive microwave data collected over Antarctica. Ann. Glaciol, 17 , 131136.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kelly, R., , Chang A. , , Tsang L. , , and Foster J. , 2003: A prototype AMSR-E global snow area and snow depth algorithm. IEEE Trans. Geosci. Remote Sens, 41 , 230242.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Knowles, K., , Njoku E. , , Armstrong R. , , and Brodzik M. J. , 2002: Nimbus-7 SMMR Pathfinder Daily EASE-Grid Brightness Temperatures. National Snow and Ice Data Center, CD-ROM.

    • Search Google Scholar
    • Export Citation
  • Latif, M., , and Barnett T. , 1994: Causes of decadal variability over the North Pacific and North America. Science, 266 , 634637.

  • Leathers, D., , and Robinson D. , 1993: The association between extremes in North American snow cover extent and United States temperatures. J. Climate, 6 , 13451355.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Loveland, T., , Reed B. , , Brown J. , , Ohlen D. , , Zhu Z. , , Yang L. , , and Merchant J. , 2000: Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data. Int. J. Remote Sens, 21 , 13031330.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • MacKay, M., , and Derksen C. , 2003: Regional snowpack modeling over Canadian landscapes. Proc. 14th Int. Symp. and Workshop on Northern Research Basins, Kangerlussuaq/Sdr. Strømfjord, Greenland, 99–105.

    • Search Google Scholar
    • Export Citation
  • MacKay, M., , Seglenieks F. , , Verseghy D. , , Soulis E. , , Snelgrove K. , , Walker A. , , and Szeto K. , 2003: Modeling Mackenzie basin surface water balance during CAGES with the Canadian Regional Climate Model. J. Hydrometeor, 4 , 748767.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Meteorological Service of Canada, 2000: Canadian Snow Data. Climate Research Branch, CD-ROM.

  • Robinson, D., 1991: Merging operational satellite and historical station snow cover data to monitor climate change. Palaeogeogr., Palaeoclimatol., Palaeoecol, 90 , 235240.

    • Search Google Scholar
    • Export Citation
  • Robinson, D., , Dewey K. , , and Heim R. , 1993: Global snow cover monitoring: An update. Bull. Amer. Meteor. Soc, 74 , 16891696.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Robinson, D., , Tarpley J. , , and Ramsay B. , 1999: Transition from NOAA weekly to daily hemispheric snow charts. Preprints, 10th Symp. on Global Change Studies, Dallas, TX, Amer. Meteor. Soc., 487– 490.

    • Search Google Scholar
    • Export Citation
  • Serreze, M., and Coauthors, 2000: Observational evidence of recent change in the northern high-latitude environment. Climatic Change, 46 , 159207.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tait, A., 1998: Estimation of snow water equivalent using passive microwave radiation data. Remote Sens. Environ, 64 , 286291.

  • Walker, A., , and Goodison B. , 2000: Challenges in determining snow water equivalent over Canada using microwave radiometry. Proc. Int. Geoscience and Remote Sensing Symp., Honolulu, HI, WE09;ll03. Institute of Electrical and Electronics Engineers, CD-ROM, WE09_03.

    • Search Google Scholar
    • Export Citation
  • Walker, A., , and Silis A. , 2002: Snow cover variations over the Mackenzie River basin from SSM/I passive microwave satellite data. Ann. Glaciol, 34 , 814.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Walsh, J., , Tucek D. , , and Peterson M. , 1982: Seasonal snow cover and short-term climatic fluctuations over the United States. Mon. Wea. Rev, 110 , 14741485.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wiesnet, D., , Ropelewski C. , , Kukla G. , , and Robinson D. , 1987: A discussion of the accuracy of NOAA satellite-derived global seasonal snow cover measurements. Large Scale Effects of Seasonal Snow Cover, B. E. Goodison, R. G. Barry, and J. Dozier, Eds., Vol. 166, International Association of Hydrological Sciences, 291–304.

    • Search Google Scholar
    • Export Citation
  • Zhang, X., , Vincent L. , , Hogg W. , , and Nitsoo A. , 2000: Temperature and precipitation trends in Canada during the 20th century. Atmos.–Ocean, 38 , 395429.

    • Crossref
    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Outline of the study domain

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 2.
Fig. 2.

(a) Average study area SWE retrievals, and relative frequency plots for the SWEunadjusted and SWEDW2003 datasets for (b) 1 Dec, (c) 1 Jan, and (d) 1 Feb 1978–87

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 3.
Fig. 3.

(a) Snow course measurement locations. (b) Frequency of passive microwave–derived SWE over- and underestimation relative to snow course measurements

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 4.
Fig. 4.

(a) Absolute difference between snow course SWE measurements and passive microwave SWE retrievals. (b) Average difference. For snow course measurement locations, refer to Fig. 3a

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 5.
Fig. 5.

Standardized anomaly sequence for the (a) SWEunadjusted and (b) SWEDW03 time series. Dashed line denotes transition from SMMR to SSM/I data. See Fig. 1 for study area domain

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 6.
Fig. 6.

(a) General classification agreement between NOAA snow extent charts and passive microwave–derived snow extent, 1978–2001. (b) Errors attributed to passive microwave omission and (c) commission

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 7.
Fig. 7.

Interannual variability in monthly SCE anomalies, standardized to a 1978–92 reference period. See Fig. 1 for study area domain

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Fig. 8.
Fig. 8.

Interannual variability in monthly SWE anomalies, standardized to a 1978–92 reference period. See Fig. 1 for study area domain

Citation: Journal of Hydrometeorology 5, 5; 10.1175/1525-7541(2004)005<0850:MCAPME>2.0.CO;2

Table 1.

Comparison of SMMR and SSM/I sensors

Table 1.
Table 2.

Correlation (r) between dataset agreement and omission and commission errors between passive-microwave-derived and NOAA snow extent data. Bold italics denote significant correlations (95%)

Table 2.
Table 3.

Correlation (r) between historical and passive-microwave-derived SCE and SWE anomalies, 1978–92. See Fig. 1 for study area domain. Bold italics denote significant correlations (95%)

Table 3.
Table 4.

Summary statistics for passive microwave and conventional SWE datasets, 1978–92

Table 4.
Table 5.

Summary of linear trend analysis results

Table 5.
Save