Location: Northwest Watershed Management Research
Project Number: 2052-13610-011-03
Start Date: Feb 01, 2013
End Date: Jan 31, 2015
We will utilize output from the Climate Forecast System (CFS), the main dynamical model used by NOAA’s Climate Prediction Center (CPC). We propose to use statistical downscaling methods to generate forecasts for relatively local-scale rangeland restoration applications. We will evaluate the accuracy of CFS forecasts with historical weather data and hindcasts for the years 1982-2010. These hindcasts include seasonal outlooks for a 9-month period subsequent to the fall planting season in the Intermountain western US. Forecast skill will be assessed using well-established statistical methods, including the Brier Skill Score, and Heidke Skill Score for variables of temperature and precipitation. We will specifically assess the skill of such forecasts for seasonal and monthly time periods associated with winter soil moisture storage, active spring germination and seedling establishment, and seedling persistence through the summer. We will also evaluate conditional skill during strong phases of the El-Nino Southern-Oscillation. This analysis will allow for the creation of skill masks that will be used with seasonal climate projections to highlight regions of greater confidence. This analysis may yield criteria for threshold forecast probabilities that trigger active management/non-management planning decisions in the fall planting season.