Data analysis Directory UMM :Data Elmu:jurnal:A:Agricultural & Forest Meterology:Vol102Issue2-3May2000:

118 D. Amarakoon et al. Agricultural and Forest Meteorology 102 2000 113–124

5. Data analysis

The data analysis involved the estimation of day- time 07:00–18:00 hours values of i 20 min and hourly latent heat fluxes, ii average evaporation for the three periods of data collection and iii evapora- tion on individual days using Eq. 6, and followed by comparisons with the corresponding measured values. The 20 min and hourly values of latent heat flux were estimated using a data set formed by combining data from all the three periods, January to February of 1994, January to February of 1990 and August to September of 1989. The individual days used in the calculations are listed in Table 3 including daytime average air temperature θ , vapor pressure deficit δe which is equal to e−e and wind speed u. The days listed in Table 3 had data for at least 8 h during daytime. Two values for the parameter α ′ in Eq. 6 were used in the calculations and β was kept at 20 W m − 2 . One value of α ′ was 0.95, which was the average value deduced by De Bruin and Holtslag 1982 using daytime hourly values of latent heat flux for their normal period, and the other was 1.05. The motivation for testing 1.05 for α ′ was the following. As stated in the study of De Bruin and Holtslag 1982, Priestley–Taylor pa- rameter α and the parameter α ′ have shown similar behavior and daily mean value of α has appeared to be about 10 greater than its hourly value during day- time. Therefore we could expect α ′ to be about 10 greater when daily values are considered, for a given β . Thus, it is justified to test the value of 1.05 for α ′ . For further evaluation of the results, we estimated i daytime evaporation on days listed in Table 3 using Table 3 Mean values of daytime temperature θ , vapor pressure deficit δe and wind speed u for selected days Day a θ ◦ C δ e mb u m s − 1 Day θ ◦ C δ e mb u m s − 1 1989227 31.4 14.1 4.7 199035 25.4 10.8 1.5 1989228 31.0 14.5 2.4 199038 27.4 14.5 2.9 1989243 31.1 11.8 3.4 199040 27.6 13.6 3.0 1989247 32.2 13.2 2.3 199041 27.3 13.7 2.9 1989248 31.6 13.0 2.2 199042 28.1 14.5 4.0 199020 30.2 21.2 2.9 199432 24.4 9.8 1.2 199021 30.0 18.4 4.8 199435 26.7 11.3 1.3 199027 29.2 17.1 2.2 199436 26.9 9.9 2.4 199033 28.2 14.8 4.4 199439 28.0 15.3 2.0 199034 29.0 17.7 5.0 199440 27.6 19.2 3.9 a Day is given as: year-day number. Penman 1948 scheme given by 1 with fu from Eq. 2 and Priestley and Taylor 1972 scheme given by Eq. 4 with 1.26 and ii daytime 20 min and hourly values of latent heat flux using Eq. 5 with α=1.26 for the combined data set. The values estimated were compared with the corresponding measured values. A value of 0.674 mb K − 1 was used for γ in the analysis. This value corresponds to p=1014 mb av- erage measured pressure, c p = 1005 J kg − 1 K − 1 and λ= 2.43×10 6 J kg − 1 Brutsaert, 1982. The values of 1 and e were estimated employing Lowe’s polyno- mials Lowe, 1977 for 1 and e, and our air temper- ature measurements.

6. Results and discussion