Open AccessUsing Geospatial Techniques and GIS to Develop Maps of Freeze Probabilities and Growing Degrees
Author(s) -
Jennifer A. Logan,
Mark Mueller
Publication year - 2000
Publication title -
hortscience
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.518
H-Index - 90
eISSN - 2327-9834
pISSN - 0018-5345
DOI - 10.21273/hortsci.35.4.558d
Subject(s) - inverse distance weighting , geospatial analysis , elevation (ballistics) , geographic information system , kriging , growing degree day , geostatistics , geography , cartography , weighting , physical geography , remote sensing , environmental science , multivariate interpolation , statistics , mathematics , spatial variability , phenology , ecology , medicine , geometry , radiology , bilinear interpolation , biology
Tennessee is located in an area of diverse topography, ranging in elevation from <100 m to ≈2000 m, with numerous hills and valleys. The physiography makes it very difficult to spatially interpolate weather data related to vegetable production, such as spring and fall freeze dates and growing degree days (GDD). In addition, there is a poor distribution of cooperative weather stations, especially those with 30 years or more of data. There are climate maps available for Tennessee, but they are of such a general format as to be useless for operational applications. This project is designed to use a geographic information system (GIS) and geospatial techniques to spatially interpolate freeze (0 °C) dates and GDD for different base temperatures and make the data available as Internet-based maps. The goal is to develop reasonable climate values for vegetable growing areas <1000 m in elevation at a 100 square km resolution. The geostatistics that we are evaluating include Thiessen polygons, triangulated irregular network (TIN), inverse distance weighting (IDW), spline, kriging, and cokriging. Data from 140 locations in and around Tennessee are used in the analysis. Incomplete data from 100 other locations are used to validate the models. GDD, which have much less year-to-year variability than freeze dates, can be successfully interpolated using inverse distance weighting (IDW) or spline techniques. Even a simple method like Thiessen produces fairly accurate maps. Freeze dates, however, are better off analyzed on an annual basis because the patterns can vary significantly from year to year. The annual maps can then be superimposed to give a better estimate of average spring and fall freeze dates.