Spectral Data-Based Estimation of Soil Heat Flux

Authors

Ramesh K. Singh, USGS EROS Center, Sioux Falls, SDFollow
Ayse Kilic, University of Nebraska-LincolnFollow
Elizabeth Walter-Shea, University of Nebraska-LincolnFollow
Shashi B. Verma, University of Nebraska-LincolnFollow
Andrew E. Suyker, University of Nebraska-LincolnFollow

ORCID IDs

Ayse Kilic

Date of this Version

2011

Citation

Transactions of the ASABE, Vol. 54(5): 1589-1597

Comments

US government work.

Abstract

Numerous existing spectral‐based soil heat flux (G) models have shown wide variation in performance for maize and soybean cropping systems in Nebraska, indicating the need for localized calibration and model development. The objectives of this article are to develop a semi‐empirical model to estimate G from a normalized difference vegetation index (NDVI) and net radiation (R_n) for maize (Zea mays L.) and soybean (Glycine max L.) fields in the Great Plains, and present the suitability of the developed model to estimate G under similar and different soil and management conditions. Soil heat fluxes measured in both irrigated and rainfed fields in eastern and south‐central Nebraska were used for model development and validation. An exponential model that uses NDVI and R_n was found to be the best to estimate G based on r² values. The effect of geographic location, crop, and water management practices were used to develop semi‐empirical models under four case studies. Each case study has the same exponential model structure but a different set of coefficients and exponents to represent the crop, soil, and management practices. Results showed that the semi‐empirical models can be used effectively for G estimation for nearby fields with similar soil properties for independent years, regardless of differences in crop type, crop rotation, and irrigation practices, provided that the crop residue from the previous year is more than 4000 kg ha^‐1. The coefficients calibrated from particular fields can be used at nearby fields in order to capture temporal variation in G. However, there is a need for further investigation of the models to account for the interaction effects of crop rotation and irrigation. Validation at an independent site having different soil and crop management practices showed the limitation of the semi‐empirical model in estimating G under different soil and environment conditions.