Snowbedo Data and Modeling Scripts

The .Rdata datasets were created to be used in the Snowbedo R Package, available on GitHub at : <a href=\"https://github.com/cahhansen/Snowbedo\" rel=\"nofollow\">https://github.com/cahhansen/Snowbedo</a>. The Snowbedo package and models were developed in part with support of the iUTAH project, under the Research Focus Area 3: Understanding the ties between human and environmental water systems.

The datasets contain a collection of time series of streamflow, meteorological, and land surface/atmospheric data which can be used to calibrate streamflow models. Each dataset corresponds to a different watershed/stream in the Wasatch Mountains.

Variables and their definitions are as follows:
Date - date, Year-Month-Day
Streamflow - streamflow rate in cms (from Salt Lake City Department of Public Utilities)
Tmax_C - maximum temperature, degrees Celsius from sub-daily measurements (SNOTEL)
Tmin_C - minimum temperature, degrees Celsius from sub-daily measurements (SNOTEL)
SWE_cm - snow water equivalent, centimeters (SNOTEL)
Albedo - average watershed albedo (derived from MOD01A1 product)
SolarRad_Whm2d - shortwave downwelling radiation, W-h/m2/day (from CERES SYN1deg product)
SnowCover - percentage of watershed covered by snow (derived from MOD01A1 product)
SnowDepth_cm - depth of snow, centimeters (SNOTEL)
Precip_cm - precipitation, centimeters/day (SNOTEL)

Additional parameters are also included for exploring the effects of the lagged parameters (lagged by one day) on streamflow.

Streams and the locations of the SLCDPU Gage Location and their corresponding SNOTEL sites are as follows:
City Creek - 40.7841, -111.883; SNOTEL Site: Louis Meadow (972)
Little Cottonwood - 40.579, -111.798; SNOTEL Site: Snowbird (766)
Lambs Creek -40.7548, -111.709; SNOTEL Site: Parley's Summit (684)
Dell Creek - 40.7809, -111.681; SNOTEL Site: Lookout Peak (596)
Big Cottonwood - 40.618, -111.780; SNOTEL Site: Mill-D (628)

The scripts are intended to be used with the Snowbedo R Package (github.com/cahhansen/Snowbedo. The scripts may be run using R Statistical Software and the dependent external packages (listed in the scripts). The Snowbedo package was developed in order to model streamflow as a result of changing snowpack dynamics (particularly albedo). The purpose of the NeuralNetwork script is to train and build a neural network model of streamflow based on climate and watershed characteristics.Results of the model are a daily time-series of streamflow covering the same time period as the input datasets. Different scenarios (with adjusted albedo) can be created with the ModelDifferencesInAlbedo.R script. The readme.txt file explains how the package can be used.

本数据集为.Rdata格式，专为Snowbedo R包开发使用，该包可在GitHub获取：<a href="https://github.com/cahhansen/Snowbedo" rel="nofollow">https://github.com/cahhansen/Snowbedo</a>。Snowbedo包及其模型的开发部分得到了iUTAH项目的支持，该项目隶属于研究重点领域3：厘清人类与自然水环境之间的关联。

本数据集集合了径流、气象以及陆面/大气数据的时间序列，可用于径流模型的校准。每个数据集对应瓦萨奇山脉（Wasatch Mountains）内的一个不同流域/溪流。

各变量及其定义如下：
- 日期（Date）：格式为年-月-日
- 径流量（Streamflow）：单位为立方米每秒（cms），数据来自盐湖城公共事业部（Salt Lake City Department of Public Utilities）
- 最高气温（Tmax_C）：单位为摄氏度，来自逐小时观测数据（SNOTEL，Snow Telemetry Network）
- 最低气温（Tmin_C）：单位为摄氏度，来自逐小时观测数据（SNOTEL，Snow Telemetry Network）
- 雪水当量（SWE_cm）：单位为厘米，数据来自SNOTEL
- 反照率（Albedo）：流域平均反照率，源自MOD01A1产品
- 下行短波辐射（SolarRad_Whm2d）：单位为瓦时每平方米每天（W·h/m²/d），数据来自CERES SYN1deg产品
- 积雪覆盖率（SnowCover）：流域积雪占比，源自MOD01A1产品
- 积雪深度（SnowDepth_cm）：单位为厘米，数据来自SNOTEL
- 降水量（Precip_cm）：单位为厘米/天，数据来自SNOTEL

数据集还包含额外参数，用于探究滞后1天的参数对径流的影响。

各溪流及SLCDPU测站位置、对应SNOTEL站点信息如下：
- 城市溪（City Creek）：坐标40.7841, -111.883；对应SNOTEL站点：Louis Meadow（站点编号972）
- 小卡顿伍德溪（Little Cottonwood）：坐标40.579, -111.798；对应SNOTEL站点：Snowbird（站点编号766）
- 兰姆溪（Lambs Creek）：坐标-40.7548, -111.709；对应SNOTEL站点：Parley's Summit（站点编号684）
- 戴尔溪（Dell Creek）：坐标40.7809, -111.681；对应SNOTEL站点：Lookout Peak（站点编号596）
- 大卡顿伍德溪（Big Cottonwood）：坐标40.618, -111.780；对应SNOTEL站点：Mill-D（站点编号628）

本配套脚本需配合Snowbedo R包（github.com/cahhansen/Snowbedo）使用，可通过R统计软件及脚本中列明的依赖外部包运行。Snowbedo包旨在基于积雪动态（尤其是反照率）变化构建径流模拟模型。NeuralNetwork脚本的功能为基于气候与流域特征，训练并构建径流神经网络模型。模型输出结果为与输入数据集时间跨度一致的日径流时间序列。可通过ModelDifferencesInAlbedo.R脚本构建不同反照率调整情景。readme.txt文件详细说明了该包的使用方法。