Premium
Effects of watershed discretization on estimation of hydrologic parameters
Author(s) -
BIAN LING
Publication year - 1997
Publication title -
transactions in gis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.721
H-Index - 63
eISSN - 1467-9671
pISSN - 1361-1682
DOI - 10.1111/j.1467-9671.1997.tb00059.x
Subject(s) - watershed , discretization , spatial analysis , spatial distribution , similarity (geometry) , mathematics , statistics , soil science , environmental science , geography , computer science , mathematical analysis , machine learning , artificial intelligence , image (mathematics)
This study evaluates how watershed discretization affects estimation of hydrologic parameters using GIS data. Two aggregation methods were evaluated using three GIS data sets for a large watershed in Kansas, which is discretized into five different levels. The two aggregation methods are weighted‐average and dominant‐value. The three GIS data sets, soils, land use, and temperature, constitute three commonly used hydrologic parameters with distinct spatial patterns. The study evaluated the aggregation effects measured in terms of statistical distribution, spatial distribution, information level, and spatial dependence of the aggregated data. Results indicate that: (1) statistically, the mean and modal values of the source data are well preserved through aggregation but with a reduced standard deviation; (2) changes in spatial patterns are less predictable than those of the statistical distribution, and the changes depend on the geometric similarity and spatial overlap between the source and target polygons; (3) the information level in general decreases with aggregation for the dominant method, and it increases for the average method although the original values are altered; and (4) spatial dependence generally increases with aggregation.