Estimation and mapping transpiration with basal and stress crop coefficients derived from remote sensing and ground-based plant water stress indicator

Abstract

The use of remote sensing for obtaining evapotranspiration (ET) from natural and agricultural surfaces is already widely used. For irrigated agriculture the two basic approaches are: (1) the solution of the energy balance equation, using remotely sensed surface temperatures and reflectances to estimate variables and components of this equation, and (2) the crop coefficient and reference evapotranspiration (ETo) approach where the crop coefficient is obtained through canopy reflectance measurements. For the latter, theoretical and field studies have shown that satellite reflectance-derived vegetation indices (VIs) are closely relate to carbon and moisture fluxes and, when combined with ground data or appropriately calibrated models, they can produce valuable estimates of crop transpiration and related processes at the canopy or ecosystem scale (D ́Urso and Calera, 2006). In this study the crop coefficient approach was used. For estimation of actual transpiration of irrigated and very high tree-density hedgerow orchards grown in Alentejo the procedure correlates (a) the basal crop transpiration coefficients (Kcb = Tmax/ETo) of the unstressed full irrigated treatment to the normalized difference vegetation index (NDVI) values assessed from Landsat5 TM and Landsat7 ETM+ (r2 = 0.86) and (b) the crop stress coefficient (Ks = Ta/KcbETo), obtained from the ratio of the sustained deficit irrigation (Ta) and fully irrigated (Tmax) daily transpiration rates, to a plant stress indicator, in the case, the basal leaf water potential (r2= 0.85). Daily tree transpiration rates on both treatments were obtained from sap flow measurements. The unstressed crop status of the full irrigation treatment was warranted from the high Willmott index of agreement (IA = 0.88) obtained with transpiration values simulated with the Penman-Monteith “big leaf” model (Willmott, 1982). In this algorithm, a specific model of bulk daily canopy conductance (Gc) for unstressed olive canopies was used (Orgaz et al., 2007). From the resulting relationship equations, known field values of leaf basal water potential and satellite-derived NDVI ́s suffice to get estimates of Ks and Kcb, respectively and from them derive and map the actual olive tree transpiration (Ta = Kcb Ks ETo) rates. The study is under way, and thus further validation applications are planned prior to using the approach for mapping olive transpiration orchards of different tree density and scale areas.