|
Statistical ModelingBecause the hydrologic cycle is a complex process that can be difficult to represent with numerical models, past history is perhaps the best way to predict what might happen in the future. NOAA, USGS, and other government agencies have long kept records of precipitation and stream flow so for many locations we know what historical peaks, and long term averages are. However, gaged locations are relatively few compared to the number of ungaged locations for which estimates must be made. Regression techniques are typically used to relate characteristics of gaged watersheds to ungaged watersheds so that information collected can be used to predict runoff for a wider area. |