Annual Snow Timing Index Rasters for the Western US and Alaska, WY2001-2019

Here, a collection of rasters describing annual snow onset (SO), snow cover duration (SCD), and day of snow disappearance (DSD) for WY2001-2019 over Alaska, Canada, and the Western United States (boundary box: -152W, -100W, 32N, 68N) are available. Units are in calendar Day of Year (DOY) for SO and DSD, and in days for SCD. Filenames follow the convention, "index_west_threshold_spatialsubdomain.tif". For example, "SCD_west_5-0000005376-0000005376.tif" indicates the file contains snow cover duration raster data obtained using a 5% threshold. Spatial subdomain naming convention is described below. These data accompany a manuscript entitled, "Investigating the Relationship Between Peak Snow-Water Equivalent and Snow Timing Indices in the Western U.S. and Alaska". A diagnostic model of peak SWE as a function of remotely sensed snow timing indices (the dataset provided here) was developed in the course of this study to address the following questions: 1) Are peak SWE and snow timing related?; 2) How does this relationship vary across the western United States?; and 3) What meteorological and topographical conditions affect this relationship? These rasters were created in the Google Earth Engine (GEE) from the MODIS MOD10A1 v006 product (MODIS/Terra Snow Cover Daily L3 Global 500m SIN Grid, Version 6). From the initial daily fSCA product, a moving median window (i.e., a low-pass filter) of filter length k=25 days was applied to obtain a binary snow cover series using fractional cover threshold of 1%, 5%, 10%, 20%, and 30% such that the timing indices may be extracted. The filter length smooths out small-scale short-lived snow deposition events that obscure snow timing. Pixels without seasonal snowpack (i.e., persistent year-round snow or little to no snow) were masked: within a given water year, pixels that 1) show no onset of snow between the 272nd day of year and end of calendar year (corresponding with the typical N. hemisphere timeframe for the start of snow accumulation), and/or 2) do not melt out between start of calendar year and the 272nd day of year (typical timeframe for the end of snowmelt) were masked. SCD was calculated as (DSD + 365 days - SO). Because of the large file size, rasters were split up spatially into multiple files by the GEE, and follow the naming convention for large file exports listed here: https://developers.google.com/earth-engine/guides/exporting. As stated, "the filename of each tile will be in the form baseFilename-yMin-xMin where xMin and yMin are the coordinates of each tile within the overall bounding box of the exported image." Each band represents a different water year and is labeled as such, e.g. "DSD_2001".本数据集包含阿拉斯加、加拿大及美国西部（边界框：西经152°、西经100°、北纬32°、北纬68°）2001至2019水文年（Water Year, WY）的雪季起始日（Snow Onset, SO）、积雪持续时长（Snow Cover Duration, SCD）与融雪结束日（Day of Snow Disappearance, DSD）栅格数据。其中，雪季起始日与融雪结束日的单位为日历日序（Day of Year, DOY），积雪持续时长的单位为天。文件命名遵循"index_west_threshold_spatialsubdomain.tif"格式，例如"SCD_west_5-0000005376-0000005376.tif"代表该文件包含采用5%阈值提取的积雪持续时长栅格数据。空间子域的命名规则详见下文。本数据集配套发表于题为"Investigating the Relationship Between Peak Snow-Water Equivalent and Snow Timing Indices in the Western U.S. and Alaska"的论文。本研究依托本数据集提供的遥感积雪时序指标，构建了峰值雪水当量（Snow-Water Equivalent, SWE）的诊断模型，旨在解答以下三个问题：1）峰值雪水当量与积雪时序是否存在关联？2）该关联在美国西部的空间分布存在何种差异？3）哪些气象与地形条件会对该关联产生影响？本数据集的栅格数据基于谷歌地球引擎（Google Earth Engine, GEE），通过MODIS MOD10A1 v006产品（MODIS/Terra每日积雪覆盖全球500米正弦格网第6版）生成。从原始每日积雪盖度分数（fractional Snow Cover Area, fSCA）产品出发，本研究采用长度为25天的移动中值窗口（即低通滤波器）进行滤波，以1%、5%、10%、20%、30%的盖度阈值提取二值积雪序列，进而获取积雪时序指标。该滤波窗口可滤除干扰积雪时序识别的小规模短时降雪事件。研究人员对无季节性积雪覆盖的像元（即常年积雪区或几乎无积雪区）进行掩膜处理：在指定水文年内，满足以下任一条件的像元将被掩膜：1）在日序第272天至年末（对应北半球典型积雪开始时段）未出现积雪；2）在年初至日序第272天（典型融雪结束时段）未完成融雪。积雪持续时长通过公式SCD = DSD + 365 - SO计算得到。由于数据量较大，谷歌地球引擎将栅格数据按空间分割为多个文件，其命名规则遵循大文件导出规范：https://developers.google.com/earth-engine/guides/exporting。如前所述，每个瓦片的文件名格式为"baseFilename-yMin-xMin"，其中"xMin"与"yMin"为该瓦片在导出影像总边界框内的坐标。每个波段对应一个不同的水文年，并以此标注，例如"DSD_2001"。