Extended Abstract (2.0M)P1.12
A station density strategy for the United States Climate Reference Network (USCRN)
Michael J. Janis, Southeast Regional Climate Center, Columbia, SC; and K. G. Hubbard and K. T. Redmond
The U.S. Climate Reference Network (USCRN) was established to improve the capacity to observe climatic change and variability across the United States. A goal of this network is to provide long-term homogeneous observations of temperature and precipitation that can be coupled with historical observations for detection and attribution of climatic change. The purpose of this study is to estimate the number and distribution of CRN observing sites required to track, within predetermined monitoring goals, the annual temperature and precipitation variability across the contiguous United States. It was hypothesized that a greater station density would be required to achieve the monitoring goal in those areas with higher spatial and temporal climate variability. The analysis was conducted by forming hypothetical networks from representative sub samples of stations in an existing higher density baseline network. This was performed on a grid cell basis to allow for the possibility of higher and lower variability in select regions. Stratified sampling techniques are used to select stations within 2.5° latitude by 3.5° longitude grid cells. Monte Carlo resampling techniques are applied to stations within each grid cell to successively lower the resolution compared to density for that grid cell in the reference or baseline network. Measures of similarity between lower resolution networks and baseline networks are generated for each grid cell. Thus, grid cell densities required to meet pre-determined network monitoring goals can be determined in a manner that reflects the climate variability from region to region. It was found that a network of 327 stations is required to meet a combined temperature-trend monitoring goal of 0.10°C per decade and a precipitation-trend monitoring goal of 2.0% of median precipitation per decade.
Poster Session 1, Posters
Monday, 10 February 2003, 2:30 PM-2:30 PM
Previous paper Next paper