Search Filter
- Publications
- Abstract of Theses and Dissertations
- Database
- Modeling and Spatial Interpolation of Potential Evapotranspiration in...
Modeling and Spatial Interpolation of Potential Evapotranspiration in Mindanao
Abstract:
A study of modeling and spatial interpolation of potential evapotranspiration in Mindanao was conducted to (1) evaluate the reliability of using modified Penman to compute potential evapotranspiration (PET), (2) develop a model on daily PET data in each station, (3) evaluate different Geographic Information System (GIS) interpolation techniques in estimating dekadal PET, and (4) generate monthly PET maps of Mindanao.
Using correlation and regression analyses, the modified Penman equation showed significant positive correlation coefficients with observed evaporation from Class A pan in selected agrometeorological stations in Mindanao and UPLB. Since the correlation coefficients ranged from 0.679 to 0.848 then the modified Penman method can be used to compute PET in Mindanao.
Time series analysis using Box and Jenkins approach was employed to contract a model for predicting daily PET values for different station in Mindanao. The study showed that first degree differencing and transformation of series removed the periodic component and achieved stability of level and variability with time.
After first differencing, nine models were developed, possessing a moving average (MA) process, and two other models showed autoregressive properties of order one. Stations modeled under integrated moving average (IMA) (0, 1, 2) showed high spatial correlation.
For regions which are served by weather station networks, one may interpolate evapotranspiration values from monitored locations to estimate similar data for other unknown areas. The estimated values may contain error due to spatial variability of PET parameters in the monitoring station. To consider this variability, GIS was used for spatial interpolation of PET.
A GIS has the capability to interpolate a full surface from a PET data at given station. The interpolation methods included inverse distance, inverse square distance, and kriging. It was found out that inverse distance method was the best interpolator of PET because it yielded the lowest root mean square error (RMSE) of 2.8, low mean absolute deviation (MAD), and highest coefficient of determination (R-square) of 0.71.
Generation of monthly PET maps for different climatic types in Mindanao was done using GIS following inverse distance interpolation method. A summary of mean monthly PET averaged over 10 years for each weather station was also generated.
It was found out that on the average the lowest PET of 3.88 mm/day was estimated for the month of December. PET increased with time until it reached its peak of 5.13 mm/day for the month of April. Monthly PET showed an increasing trend from December.
During the month of December, low PET values (3.5-4.0 mm/day) were estimated for the whole of Mindanao except in some isolated regions in Davao and Zamboanga Cities. Estimated PET values tend to decrease as elevation increase such as in Marawi and Malaybalay.
During the months of April and May, higher PET values (5-6 mm/day) were computed for the eastern part of Mindanao. The observed decrease in PET with decreasing longitude could possibly be due to trade winds from the Pacific Ocean.
PET maps showed that there are four existing groupings, indicating that there are no pronounced wet and dry seasons in Mindanao.