Assessing the Agricultural Reference Index for Drought (ARID) Using Uncertainty and Sensitivity Analyses

The Agricultural Reference Index for Drought (ARID) is a newly designed decision support tool to quantify plant water deficit and predict the effect of deficit on crop yields. This study explored uncertainties in ARID associated with its parameters and the sensitivity of ARID to its parameters. Daily values of ARID were computed for five selected locations in the southeastern United States using historical weather data for a 30-yr period (1971-2000). Uncertainty and sensitivity analyses were performed using the Fourier amplitude sensitivity test. According to the results, available water capacity was the most influential parameter, explaining about 60% of the total variance in ARID, followed by root zone depth, which contributed about 30% to the total variance. Of the five parameters, runoff curve number and drainage coefficient had insignificant influence. The effect of the water uptake coefficient was negligible in most cases, and this parameter never contributed more than 15% of the total variance. Except for soils with large moisture content, ARID had small uncertainties associated with its parameters. Even under wet conditions, ARID mostly concentrated around its default values, indicating small uncertainties, which were mainly due to available water capacity. The type of distribution selected for the parameters was found to have significant influence on sensitivity and uncertainty. With a shift from uniform to normal distribution, uncertainty in ARID decreased by 15 to 50%. Results indicated that although ARID uses a fixed set of parameter values, it is applicable to a wide range of crops, soils, topographies, and management and has fairly small uncertainties.