Spatial modeling of evapotranspiration in the Luquillo Experimental Forest of Puerto Rico using Remotely-sensed data.

Actual evapotranspiration (aET) and related processes in tropical forests can explain 70% of the lateral global energy transport through latent heat, and therefore are very important in the redistribution of water on the Earth’s surface [Mauser, M., Schädlich, S., 1998. Modelling the spatial distribution of evapotranspiration on different scales using remote sensing data. J. Hydrol. 212–213, 250–267]. Unfortunately, there are few spatial studies of these processes in tropical forests. This research integrates one Landsat Thematic Mapper (TM) image and three Moderate Resolution Imaging Spectroradiometer (MODIS) images with a hydrological model [Granger, R.J., Gray, D.M., 1989. Evaporation from natural nonsaturated surfaces. J. Hydrol. 111, 21–29] to estimate the spatial pattern of aET over the Luquillo Experimental Forest (LEF) – a tropical forest in northeastern Puerto Rico – for the month of January, the only month that these remotely sensed images were acquired. The derived aETs ranged from 0 to 7.22 mm/day with a mean of 3.08 ± 1.35 mm/day which were comparable to other estimates. Simulated aET was highest in the low elevation forest and decreased progressively toward higher elevations. Because of differences in solar radiation at different elevations, aspects and topographic positions, aET tended to be higher on south slopes and along ridges than on north slopes and in valleys. In addition, the Bowen ratio (the ratio of sensible heat to latent heat) varied across different vegetation types and increased with elevation, thus reflecting differences in the distribution of net solar radiation incident on the earth’s surface. Over a day, the highest simulated aET occurred at around noon. We also applied this model to simulate the average monthly aET over an entire year based on the cloud patterns derived from at least two MODIS images for each month. The highest simulated aET occurred in February and March and the lowest in May. These observations are consistent with long term data. The simulated values were compared with field measurements of the sap flow velocity of trees at different elevations and in different forest types. These comparisons had good agreement in the low elevation forest but only moderate agreement in the elfin forest at high elevations.