Modeling Future Climate for Model My Watershed

This resource demonstrates the workflow developed to prepare downscaled GCM data for input to Model My Watershed (ModelMyWatershed.org). GCM data for the Delaware River Basin was assembled from 19 GCMs including each model's RCP4.5 and RCP8.5; this was performed by Dr. Tim Hawkins, Shippensburg University (http://www.ship.edu/geo-ess/). Downscaled precipitation data from global climate models (GCM) does not accurately retain the magnitude and frequency of individual storm events for a given location. This lack of predictive resolution of event magnitude and frequency limits realism of rainfall-runoff models used to for predicting watershed hydrology under future climate scenarios. To address this problem, Maimone et al (2019) developed a method for summarizing the statistical distribution of precipitation event magnitude and frequency that could be applied to downscaled GCM precipitation predictions. Application of the methods here to down-scaled GCM scenarios requires that the those predictions do not include an increase in the number of days of precipitation per year. Maimone et al (2019) state this requirement: \"Because GCM projections for the Philadelphia region do not indicate an increase in the number of wet days per year, future increases in precipitation are the result of the existing number and distribution of wet days becoming more intense.\"

I developed a workflow to replicate Maimone et al's methods and provide an example of it in this Resource. There are three sections of the R Markdown document. The first section seeks to replicate the synthetic weather generator developed by Maimone et al (2019) using an example dataset. The second section applies those methods to the downscaled GCM ensemble average conditions for the Delaware River Basin provided by Dr. Hawkins. The third section develops depth-duration-frequency statistics for the 24 hour storm event relevant to the 2080-2100 predictions. To open the R Markdown document and execute the workflow yourself, find the Open With dropdown list in the upper right hand corner of this Resource and select CUAHSI JupyterHub.

The first section uses an example precipitation dataset from the Philadelphia Airport for the period 01 January 1995 through 31 December 2013. The data were downloaded from NOAA's Climate Data Online Search portal: https://www.ncdc.noaa.gov/cdo-web/search.

The downloaded data and metadata for this NOAA Climate Data are available on Hydroshare here: http://www.hydroshare.org/resource/60058ceda8334e68be141516c5b8de3f.
Additional data on precipitation frequency at the Philadelphia Airport was downloaded from the NOAA Hydrometeorological Design Studies Center: https://hdsc.nws.noaa.gov/hdsc/pfds/index.html.

An example of working with this type of NOAA Climate Data is provided on the NEON website here:
https://www.neonscience.org/da-viz-coop-precip-data-R.

References:
Maimone, M., S. Malter, J. Rockwell, and V. Raj. 2019. Transforming Global Climate Model Precipitation Output for Use in Urban Stormwater Applications. Journal of Water Resources Planning and Management 145:04019021.

本数据集展示了为流域模型在线平台（ModelMyWatershed.org）输入准备降尺度全球气候模型（Global Climate Model, GCM）数据的开发流程。特拉华河流域的GCM数据由19个GCM模型及其典型浓度路径（Representative Concentration Pathway, RCP）4.5与RCP8.5情景拼接而成，该数据整理工作由什普彭斯堡大学的蒂姆·霍金斯博士完成（http://www.ship.edu/geo-ess/）。

全球气候模型降尺度后的降水数据无法准确保留特定区域单次暴雨事件的量级与发生频率，这种对事件量级和频率预测分辨率的不足，会降低未来气候情景下水文预测所用降雨-径流模型的真实性。为解决这一问题，Maimone等人（2019）提出了一种可应用于降尺度GCM降水预测的方法，用于总结降水事件量级与频率的统计分布。

需注意，将本文方法应用于降尺度GCM情景时，需满足该预测未包含年降水日数增加的前提。Maimone等人（2019）对此作出说明："由于费城区域的GCM预测未显示年降雨日数增加，未来降水的增加源于现有降雨日的数量与分布强度提升。"

本人开发了一套复现Maimone等人方法的流程，并在本数据集中提供示例。该R语言标记文档（R Markdown）包含三个部分：第一部分借助示例数据集复现Maimone等人（2019）开发的合成天气发生器（synthetic weather generator）；第二部分将上述方法应用于霍金斯博士提供的特拉华河流域降尺度GCM集合平均（ensemble average）情景；第三部分针对2080-2100年的预测结果，生成与24小时暴雨事件相关的雨量-历时-频率（Depth-Duration-Frequency）统计量。如需打开该R Markdown文档并自行执行流程，请点击本数据集右上角的"打开方式"下拉列表，选择CUAHSI JupyterHub。

第一部分使用了1995年1月1日至2013年12月31日期间费城国际机场的降水示例数据集，该数据从美国国家海洋和大气管理局（National Oceanic and Atmospheric Administration, NOAA）的气候数据在线搜索门户下载（https://www.ncdc.noaa.gov/cdo-web/search）。

该NOAA气候数据的下载文件与元数据可在Hydroshare平台获取：http://www.hydroshare.org/resource/60058ceda8334e68be141516c5b8de3f。此外，费城国际机场的降水频率补充数据从NOAA水文气象设计研究中心下载：https://hdsc.nws.noaa.gov/hdsc/pfds/index.html。

针对此类NOAA气候数据的处理示例可在NEON官网获取：https://www.neonscience.org/da-viz-coop-precip-data-R。

参考文献：
Maimone, M., S. Malter, J. Rockwell, and V. Raj. 2019. Transforming Global Climate Model Precipitation Output for Use in Urban Stormwater Applications. *Journal of Water Resources Planning and Management* 145:04019021.