Evapotranspiration and Crop Coefficients for Seepage-Irrigated Watermelon with Plastic Mulch in a Sub-Tropical Region

Abstract

<abstract><title><italic>Abstract.</italic></title> The crop coefficient (K_c) method is the most widely used method for computing crop evapotranspiration (ET_c) and relies on selecting a representative K_c value for a specific production system, soil, and climate. Literature-based (e.g., FAO-56) K_c values, used in the absence of local K_c, can lead to erroneous ET_c estimates. This error becomes considerably high when the standard K_c values from the open-field production systems are applied to plastic mulch production systems with seepage irrigation that results in shallow water table conditions. A three-year study (2003-2005) was conducted with two large drainage lysimeters (length = 4.87 m, width = 3.65 m, and depth = 1.37 m) to quantify ET_c and K_c for seepage-irrigated watermelon under plastic mulch in sub-tropical Florida. The seasonal ET_c varied from 344 to 422 mm, with an average of 373 mm. The K_c values for initial, development, mid-season, and late stages were 0.64, 1.00, 1.28, and 1.15, respectively, and were statistically higher than the values reported in the literature, including FAO-56. The initial K_c especially was greater than the literature values due to low plant cover and high surface soil moisture in the row-middle areas at the beginning of the season resulting from frequent rainfall and shallow water table. To adjust K_c for the effect of variable rainfall wetting on evaporation during the initial stage, a multivariate regression model (r² = 0.74) was developed as a function of rainfall and solar radiation. Use of literature-based K_c values resulted in underestimation of seepage-irrigated ET_c by 26% to 43%; the 26% underestimation was from a comparison using the same lysimeter setup and seasons but with drip (surface) irrigation. Such large differences, not reported earlier, highlight the importance of developing local K_c for a specific irrigation method, climate, and production system. Regression models (r² = 0.96) were developed for predicting K_c as functions of time and growing degree days. With approximately 65,000 ha of vegetables grown under seepage irrigation in Florida, there is a need for developing local K_c values for irrigation management and water allocations. Scaling up the errors in water use from literature-based K_c values to Florida’s seepage-irrigated vegetable lands for one season per year showed errors of 25 to 42 million m³ of water. This error accounts for 15% to 25% of annual rainfall and is likely to be much higher for dual cropping systems within a year. The extent of errors and related effects on watershed water balances highlight the need for similar studies for other seepage-irrigated mulched crops.