PAPER TITLE :EMPIRICAL MODELS FOR THE PREDICTION OF GLOBAL SOLAR RADIATION IN ARID AND COASTAL/RAINFOREST REGIONS OF NIGERIA

FUTA JOURNAL OF RESEARCH IN SCIENCE | VOLUME 10 NUMBER 1 2014

Paper Details

  • Author(s) : E. O. Ogolo*1 and O. B. Daramola2
  • Abstract:

In this study eight empirical models were developed for the simulation of global solar radiation (GSR) (both linear
and multilinear) using multilinear regression technique for nine stations drawn from two regions of Nigeria; namely
Arid and Coastal/Rainfall regions. The developed empirical models as functions of meteorological variables are
M1 (sunshine hour), M2 (cloud cover), M3 (precipitation), M4 (sunshine hour and cloud cover), M5 (sunshine hour
and precipitation), M6 (sunshine hour, relative humidity and cloud cover), M7 (sunshine hour, relative humidity
and precipitation) and M8 (sunshine hour, air temperature, relative humidity, cloud cover and precipitation) for the
simulation of GSR. The performance and the efficiency of the developed models in terms of the simulated and
observed GSR was carried out, using some statistical indicators which include coefficient of determination (R2),
mean bias error (MBE), root mean square error (RMSE) and standard error of estimate (SEE). Generally, all the
models show the tendency of overestimation as the calculated MBE was positive for both regions. In Arid region,
the highest coefficient of determination (R2 = 0.61) was recorded by model 4, a sunshine hour and cloud cover
based model, while model 8 which combined all the parameters recorded the lowest MBE(2.08), RMSE(3.08) and
SEE(1.83). For the Coastal/Rainfall region, model 8 is found to be more suitable compared with other empirical
model as it has the highest R2 (= 0.74), lowest RMSE (= 1.47), MBE (0.80) and SEE (1.25). It is further observed
that while all the sunshine hour and precipitation based models (M3, M5, M7 and M8) are characterized with
tendency to perform better when compared with the sunshine hour and cloud cover based models (M2, M4 and
M6) in Coastal/Rainforest region the reverse is the case in Arid region of the country . It was also discovered that
the two schemes found to exhibit best performance for each region also perform better than Angstrom-Prescott and
are hence recommended for the estimation of GSR in the arid and coastal regions of Nigeria