Soil moisture and the water balance in a border-irrigated field

Abstract

Sampling and analysis of the soil moisture distribution and the overall water balance in an irrigated area are the central topics of this work. An experimental study was made in a 14-ha, border-irrigated, alfalfa field near Coolidge, in Final County, Arizona, during the summer/fall 1983. The water stored in the soil profile and its change with time were normally distributed, with coefficients of variation of about 10 and 25 percent, respectively. Temporal correlations were significant for storage (about .60), but absent in the other. Variograms were calculated to show the spatial structure of the distributions. An analogous statistical description was presented for the alfalfa yield. Also shown is a methodology to infer errors due to the field calibration of the neutron probe. Another task was to assess a methodology to minimize sample numbers for soil-water storage. Following the ideas of Vachaud and co-workers in France, it was verified that rankings of the measurements were approximately time-preserved. As a consequence, only a few key locations need to be sampled to evaluate the mean in those circumstances. Included also is an approximation to predict confidence intervals for estimating the mean, when such "representative sites" are used. Using irrigation inflow and rainfall, a procedure is defined to make use of the soil moisture data to evaluate irrigation efficiency and uniformity. Evapotranspiration (ET) distribution can also be assessed by soil moisture measurements, but only conditionally. For example, adaptation of the "field capacity" concept in the field study led to average daily ET rates in the range of 3-11 mm day⁻¹. ET and potential ET (PET) were also determined from weather data. Crop temperature was required in the ET calculation. Such a model, developed by Hatfield and co-workers, was judged satisfactory in our application, but not the Penman PET estimates. It is concluded that the ET model is promising, particularly if remote sensing of the temperatures is successful in the future. Also shown as a possibility is the use of plant temperature and pan evaporation data to infer crop water stress.