DOWNLOADING DATA VIA HTTPS
To learn more about Earthdata Login and register for an account, please see How to Register with Earthdata Login. Once you have logged in, data can be downloaded via a Web browser, command line, or client. For help with downloading data, please see Options Available for Bulk Downloading Data from HTTPS with Earthdata Login.
Canadian Meteorological Centre (CMC) Daily Snow Depth Analysis Data, Version 1
This data set consists of a Northern Hemisphere subset of the Canadian Meteorological Centre (CMC) operational global daily snow depth analysis. Data include daily analysed snow depths from 1998 through 2016, snow depth and estimated snow water equivalent (SWE) monthly averages, plus monthly climatologies of snow depth and estimated SWE. Please note that this data set is NOT homogeneous. It is derived from operational data that is subject to frequent changes. See the data set documentation for details.
|Temporal Resolution:||1 day|
|Sensor(s):||SNOW MEASURING ROD|
|Data Contributor(s):||Ross Brown|
|Metadata XML:||View Metadata Record|
As a condition of using these data, you must cite the use of this data set using the following citation. For more information, see our Use and Copyright Web page.Brown, R. D. and B. Brasnett. 2010, updated annually. Canadian Meteorological Centre (CMC) Daily Snow Depth Analysis Data, Version 1. [Indicate subset used]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: http://dx.doi.org/10.5067/W9FOYWH0EQZ3. [Date Accessed].
Detailed Data Description
This data set contains daily-analysed snow depths from 1998 through 2016, snow depth and estimated SWE monthly averages, plus monthly climatologies of snow depth and estimated SWE through 2012. The snow depth analysis is generated from real-time, in situ daily snow depth observations, using optimal interpolation with a first-guess field generated from a simple snow accumulation and melt model driven with analyzed temperatures and forecast precipitation from the Canadian forecast model (see Brasnett, 1999). In situ observations include snow depths from surface synoptic (synop) observations, meteorological aviation (metar) reports, and special aviation (SA) reports from the World Meteorological Organization (WMO) information system.
Note: This data set is NOT homogeneous. It is derived from operational data that is subject to frequent changes. See the Warnings and Notices section of this document for details
Data files are provided in compressed (.zip) ASCII tab-delimited format. In addition, a comma-separated value data mask is available for continental-scale studies (see Warnings and Notices: #2).
NOTE: December 15, 2016: GeoTIFF versions of the daily snow depth files for the 1998-2016 period have kindly been made available to the research community by Robert Way (U. Ottawa) using scripts developed by Dr. Koen Hufkens (Harvard University). Please contact Ross Brown (email@example.com; firstname.lastname@example.org) to get ftp access to the GeoTIFF files.
Northern Hemisphere Daily Snow Depth Files
A Northern Hemisphere subset of the CMC daily global analysis was interpolated to a standard 24 km polar stereographic grid that closely approximates the grid used by NOAA for the 24 km daily IMS snow product. At 706 pixels, the grid covers a slightly smaller area than the full 1024-pixel IMS grid. The land/sea mask should be used to remove snow over sea ice and over Greenland.
North Polar Stereographic
ni (# rows) = 706, nj (# cols) = 706, bottom left grid point = (1,1)
Grid rotation -10°
Pole position = (353, 353)
Grid resolution at 60°N = 23812.5 m
Latitude and longitude coordinates corresponding to the center of each grid cell are stored in cmc_analysis_ps_lat_long.txt (compressed as cmc_analysis_ps_lat_long.zip on the FTP site). This file lists each cell and its geographic coordinates as a row, column (i, j) ordered pair starting with (1,1), the cell in the lower left corner of the polar stereographic grid. Cell locations then increment across each column of the first row, each column of the second row, and so on, until reaching the upper right corner at (706,706). Refer to Figure 1 to see the first 10 entries from cmc_analysis_ps_lat_long.txt:
Figure 1. Latitude and longitude coordinates for cells (1, 1) through (1, 10) from cmc_analysis_ps_lat_long.txt
Each file contains one year of data with the following FORTRAN statements:
do 1001 iday=1,ndays ! 365 or 366 depending on # days in year write(60,*) idate ! date in format YYYYMMDDHH where HH is always 00 do 1002 i=1,NI write(60,100) (sdep(i,j),j=1,706) !cm 100 format(706(1x,f5.1)) !values exceeding 1000 set to 999.9 1002 continue 1001 continue
Note: In 1998, the data start at 01 August, listed as 1998080100 in the data file.
The CMC analysis includes snow on sea ice. A custom binary (0,1) land/sea mask is provided to mask out ocean areas and Greenland. Land (mask=1) was assigned for grid land cover fractions greater than or equal to fifty percent. The array should be read in as follows:
open(50,file='cmc_analysis_lsmask_binary_nogland.txt',status='old') do 5 i=1,706 read(50,21) (lsmask(i,j),j=1,706) 21 format(706(i1)) 5 continue
Monthly Average Snow Depth Files
Monthly average snow depth data are included with one year of data per file with naming convention: cmc_analysis_mly_avg_YYYY.txt. An upper snow depth limit of 600 cm was applied to the recent analyses to maintain continuity. The land/sea mask should be used to remove snow over sea ice and over Greenland where the analysis is unreliable.
Each file contains one year of data with the following FORTRAN statements:
do 1000 imon=1,12 write(60,*) iyear,imon do 1001 i=1,ni write(60,100) (sd_avg(i,j),j=1,706) !cm 100 format(706(1x,f5.1)) 1001 continue 1000 continue
Note: In 1998, the data start in August.
The directory structure on the HTTPS site is described in Table 1.
|Top Level Directory||Contents|
|nsidc0447_CMC_snow_depth_v01||Along with the three subdirectories below, the top-level directory contains the following files:
|Monthly_Climatologies¹||Contains the following files:
|Monthly_Snow_Depth_Files||Contains monthly average snow depth files for each calendar year from 1998–2016:
|Monthly_SWE_Estimates||Contains one zip file with monthly average SWE values for each calendar year from 1998–2016:
|¹Climatologies are not updated beyond 2012 to provide a fixed reference period for anomaly tracking.|
For a description of file names, refer to Table2.
The data files employ a variety of naming conventions. All file name variables are listed and described in Table 2.
The files included in this data set range from approximately 1 MB to 900 MB.
The volume of this data set is approximately four gigabytes.
Southernmost Latitude: 0° N
Northernmost Latitude: 90° N
Westernmost Longitude: 180° W
Easternmost Longitude: 180° E
The analysis has a resolution of 1/3° Gaussian. The error correlations used by the analysis have an e-folding distance of approximately 120 km horizontally and 800 m vertically.
Data include daily observations from 01 August 1998 to 31 December 2016. Note: the daily analysis for 08 December 2014 was missing, so the investigators replaced it with the previous day.
The analysis is updated every six hours, but only the 00 hour analysis is archived in this data set. For more information, refer to the Processing Steps section of this document.
This data set includes all available real-time snow depth observation measurements from synops, metars, and SAs.
Snow Water Equivalent
Monthly average estimates of SWE (mm) are also included in this data set. Monthly average SWE for the October through June period of each year was estimated from monthly average snow depth files by applying a look-up table (see Table 3) of mean monthly snow density values. Mean monthly snow density values were derived from Canadian snow course observations corresponding to snow climate classes in the Sturm et al. (1995) classification. The July through September time period was not processed due to a lack of observed snow density information.
Table 3 lists the mean monthly snow density look-up table used to estimate SWE based on Canadian snow course observations averaged over the snow-climate classes defined by Sturm et al. (1995) (Brown and Mote, 2009). This assumes the Canadian density observations are representative of snow-climate classes in other regions of the Northern Hemisphere.
The Sturm et al. (1995) snow climate classification was interpolated from the 0.5° by 0.5° gridded version (ARCSS045) at the University Corporation for Atmospheric Research (UCAR) (Liston and Sturm 1998) to the 24 km polar stereographic grid of the monthly averaged snow depth. This process resulted in land/sea confusion at a number of points, which were excluded from the analysis. The number of points affected by this problem (nconf) is indicated in the header record for each month. The SWE estimates are contained in a single file (cmc_mly_swe_1998to2016.txt) with -999.0 values at all grid points where SWE could not be calculated (such as ocean or Sturm snow-class undefined). Greenland was excluded from the analysis, as the analysis is considered unreliable in this area. Values were written as tab-delimited with the following FORTRAN statements: write(60,*)iyear,imonth,nconf do 1002 i=1,ni write(60,107) (swe(i,j),j=1,706) !mm 107 format(706(1x,f6.1)) 1002 continue Note: In 1998, values start in October.
Snow Depth and Estimated SWE Climatologies
Monthly average snow depth for the 1998/99 to 2013/14 snow seasons are included along with an animation of monthly mean snow depth averaged over the 1999-2006 period (cmc_monthly_mean_sdep_animation_1999-2006.avi). The land/sea mask was applied to these outputs. A legend for the snow depth animation is included in a separate GIF image file (units are cm). SWE climatologies are provided for the 1998/99 to 2012/13 snow seasons.
The monthly snow depth climatology includes values from January to December, while the SWE climatology includes values from October to June. Both files are written in comma-delimited latitude-longitude tagged format with a single header record containing the variable names. Climatologies are not updated beyond 2012 to provide a fixed reference period for anomaly tracking.
Note: Starting in 2013/2014, processing was migrated from a UNIX platform to a PC. As such, zero values in some files after this date are written as .0 instead of 0.0.
Sample Analysis Image
Figure 2 displays an example of the CMC analysis data used for evaluating Global Climate Model (GCM) output from Brown and Mote (2009).
Software and Tools
Tools appropriate for viewing these data include any video or media software package and any text editor or Web browser.
A complete description of the analysis methodology and validation is published in Brasnett (1999). The methodology was applied to carry out a reanalysis of snow depths over North America for the Atmospheric Model Intercomparison Project - 2 (AMIP-2) project from 1979-1997 using a more elaborate snow model for the first guess field (Brown, et al. 2003). The scheme makes use of an analysis of screen-level temperature produced with vertical correlations taken into account.
Over most of the Arctic region, there are no observations, so the analysis is based on estimated snow depths from the first guess field. In addition, snow depth observations over northern Canada tend to be biased to coastal locations with observing sites at open areas near airports. The snow at these sites tends to be shallower and to melt out earlier than snow in surrounding terrain.
In cold, high precipitation environments, such as Greenland, the snow accumulation is capped at 600 cm to avoid excessive amounts of snow accumulating. The analysis is not considered reliable in these regions. This limit was changed to 1200 cm in October 2008 when a Limited Area Model (LAM) version of the analysis was implemented over British Columbia. Verification of that product showed it had insufficient residual snow during the summer months due to the 600 cm limit, so the limit was also changed in the global analysis. This limit is not unreasonable; mean annual snowfall in some locations in British Columbia exceeds 1000 cm, as illustrated by the mean annual snowfall levels for the 1951 through 1980 normal climate period listed in Table 4:
|Allison Pass||1431.5 cm|
|Rogers Pass||1106.7 cm|
|Pine Pass||1075.5 cm|
|Tahtsa Lake||1041.2 cm|
|Mt. Fidelity||1974.8 cm|
Values exceeding 1000 cm after 01 January 2007 were set to 999.9 to conform to the existing database format.
Data Acquisition and Processing
Real-time snow depth data were originally acquired from the World Meteorological Organization (WMO) information system.
The analysis is updated every six hours using the method of optimum interpolation with an initial guess field provided by a simple snow accumulation and melt model using analyzed temperatures and forecast (six hour) precipitation from the CMC Global Environmental Multiscale (GEM) forecast model. The precipitation is assumed to be snow if the analyzed screen-level temperature is less than 0 degrees. A degree-day melting algorithm removes mass from the snowpack at the rate of 0.15 mm-h-1-K-1.
In regions where there are no observations of snow depth, the snow depth shown in the analysis corresponds to the initial guess field simplified assumptions regarding snowfall, melt and aging.
Note: When there is an ice cover, the analysis will accumulate snow on it so the snowfield can extend over water bodies. This can be masked out if desired.
The first-guess snow model assumes the density of new snow to be 100 kg/m3 and the snowpack gradually increases in density as it ages. The increase in density with aging stops when the density reaches 300kg/m3 (except 210 kg/m3 if the vegetation is needleleaf forest, since in these regions, the canopy shelters the snowpack from wind and sunlight and densities are less). New snow causes the density to decrease by an amount related to the mass of the new snow and the mass of the existing snowpack. During melting, the density is allowed to increase up to a maximum of 550 kg/m3.
The analysis had been run essentially unchanged from 12 March 1998. An error was introduced into the operational implementation of the analysis in October 2006, but this has been corrected in offline runs and an offline version of the analysis is being run in parallel with the operational analysis to maintain data continuity.
- Prior to 04 January 2010, the analyses from October 2006 were inadvertently run with the vertical correlation component of the optimal interpolation switched off which would have affected snow depths in mountainous areas. These files were corrected and replaced on 04 January 2010. The error resulted in slightly lower continental snow cover extent (less than 10 percent) from October to June, but 10 to 20 percent underestimates of snow cover extent from July to September.
- The snow depth analysis is NOT homogeneous in data sparse regions! A change of resolutions from 100 km to 33 km on October 31, 2006 (and the removal of an upper limit cap on snow accumulation) to the Global Environmental Multiscale (GEM) model generating the forecast precipitation fields has resulted in increased SWE values. A preliminary evaluation over Arctic land areas suggests maximum SWE values are about 20 mm higher after the implementation of the higher resolution version of GEM. The impact is even more noticeable in high SWE regions such as mountains, which exhibit an increasing trend in annual maximum SWE since 2007. In continental-scale studies, the increasing trend is reduced if regions with mean annual maximum accumulations exceeding 300 mm are masked out (see Figure 3). A more aggressive mask (cmc_homog_mask_points.csv, available on the FTP site) with the location of points to exclude is needed to maintain homogeneity in continental-scale studies in recent years, due to a further increase in the resolution of GEM forecast precipitation to 25 km in 2013.
- Over most of the Arctic and mountain regions, there are no observations, so the analysis is based essentially on estimated snow depths from the first guess field. In addition, snow depth observations over northern Canada tend to be biased to coastal locations with observing sites at open areas near airports. The snow at these sites tends to be shallower and to melt out earlier than snow in surrounding terrain.
- Data prior to August 1998 have been removed from the data archive, as the snowpack model used in the analysis was unable to be properly initialized for the 1997/98 snow season. The snow depths have been determined to be anomalously high in consequence.
- An improved land/sea mask has been generated for use with the analysis.
- Snow depth and SWE values over Arctic land areas in March, April and May 2002 are anomalously low compared to other data sets. There is evidence from Canadian in situ observations that spring 2002 snow depths were the shallowest observed since 1951. Nevertheless, the striking discrepancy between CMC and other data sets in these particular months suggests that it may be prudent to exclude them from any analysis.
- The daily analysis for 08 December 2014 was missing. To fill this gap, the investigators replaced it with the previous day's data.
- Starting in 2013/2014, processing was migrated from a UNIX platform to a PC. As such, zero values in some files after this date are written as .0 instead of 0.0.
- The current product is scheduled for phase-out in 2015–2016 when the new Canadian Land Data Assimilation System (CaLDAS) system becomes operational. See Carrera et al., 2015.
- Harry Stern at the Polar Science Center, University of Washington, documented an anomalously large decrease in snow depths around 10 December 2009, with a sharp increase on 01 January 2010 at several locations on Baffin Island. This problem was found to affect all mountain regions with high snow accumulation (areas in dark blue shown in Fig. 4). The precise reasons for this dropout have not been identified.
- Michael Notaro at the Nelson Institute for Environmental Studies' Center for Climatic Research, University of Wisconsin-Madison, documented instances where the CMC analysis substantially underestimates snow depths in areas bordering Lake Superior in northern Michigan. In this region, CMC annual maximum snow depths are about half the observed values. A lack of real-time snow depth observations and under-represented lake-effect snowfall in the CMC precipitation forecast contribute to this underestimate. In this area, snow depths from the Snow Data Assimilation System (SNODAS) are likely more realistic as SNODAS has access to more real-time observations of snow depth than the CMC global analysis.
- Grant of License:
The Government of Canada (Environment Canada) is the owner of all intellectual property rights (including copyright) of this Data Product. You are granted a limited, non-exclusive, non-assignable and non-transferable license to use this data product for research purposes only subject to the terms below. This license is not a sale of any or all of the owner's rights. This product may only be used by you, and you may not rent, lease, lend, sell, sub-license or transfer the data product or any of your rights under this agreement to anyone else.
- Redistribution Restrictions:
You are not authorized to distribute the data further, including any portion of it, contained in this product.
- No Warranties:
Environment Canada does not warrant the quality, accuracy, or completeness of any information or data. Such information and data is provided "as is" without warranty or condition of any nature. Environment Canada disclaims all other warranties, expressed or implied, including but not limited to implied warranties of merchantability and fitness for a particular purpose, with respect to the software, the data retrieved from this product, and any accompanying materials.
- Restriction and Limitation of Liability:
In no event shall Environment Canada be liable for any other damages whatsoever (including, without limitation, damages for loss of business profits, business interruption, loss of business information, or other pecuniary loss) arising out of the use of, or inability to use this Environment Canada product, even if Environment Canada has been advised of the possibility of such damages.
- Responsible Use:
It is your responsibility to ensure that your use of this product complies with these terms and to seek prior written permission from Environment Canada and pay any additional fees or royalties, as may be required, for any uses not permitted or not specified in this agreement.
- Acceptance of this Agreement:
Any use whatsoever of this Data Product shall constitute your acceptance of the terms of this agreement.
References and Related Publications
Contacts and Acknowledgments
Ross D. Brown
Climate Processes and Earth Observation Division
Environment Canada @ Ouranos
550 Sherbrooke St. West, 19th Floor
Montréal QC H3A 1B9
The PI wishes to thank Bruce Brasnett, Meteorological Service of Canada, for his unfailing help and support in providing this dataset to the scientific community.
DOCUMENT CREATION DATE
DOCUMENT REVISION DATE