Geology Reference
In-Depth Information
Fig. 5.18 Observed versus LLR modelled solar radiation in the testing data set for scenario 2
5.4 Data Based Daily Solar Radiation on Beas Database
In the previous section, we have tackled some data modelling issues with GT,
Entropy Theory, BIC and AIC using the 6-hourly records from the Brue catchment.
It was found that the performance of model selection tools such as AIC/BIC weren
t
that encouraging compared to that of entropy theory. On the other hand, it is very
encouraging to
'
find that the new GT was able to choose the best input selection
during the testing with the non-linear models LLR and ANN. To emphasis more on
the reliability of GT and its applicability with other nonlinear models we have
explored further using the daily data from the Brue catchment and are presented in
this section.
The HYREX data set from the Brue catchment was used for the analysis.
A major problem associated with the HYREX during the study period was data
discontinuity. We managed to collect a total of 1,098 daily records from 1993 to
1996 after the missing data were taken out. All the data were normalised prior to
analysis by mapping the mean to zero and the standard deviation to 0.5. The process
of normalisation of data helps to keep the asymptotic nature of the Gamma Statistic
valid. In this part of the study, daily information of the following input parameters
have been considered: horizontal extraterrestrial radiation (based on Allen et al.
[
8
]), mean air temperature (averaged over 24 h), maximum daily air temperature,
minimum daily air temperature, wind speed (averaged over 24 h), and rainfall
(summed over 24 h). The daily extraterrestrial radiation can be estimated from the
solar constant, the solar declination and the time of the year Allen et al. [
8
].
