|
|
|
Abstract
Spatial Interpolation Techniques for Deviation of Mean Daily Temperature from Normal
Fariza Yunus
Asst. Director, Training Division
Malaysian Meteorological Department,
Malaysia
Email: fariza@kjc.gov.my
Radzali Mispan
Malaysian Agriculture Research and Development Institute
Ahmad Rodzi Mahmud
Geomatic Unit
Universiti Putra Malaysia
In Malaysia where temperature is almost uniform, deviation of mean daily temperature from normal is more appropriate to represent mean daily temperature at a specific area. There are several interpolation techniques to generate spatial distribution for deviation of mean daily temperature from normal. Two interpolation techniques were examined, namely Ordinary Kriging (OK) and Inverse Distance Weighting (IDW). For each of this technique, three interpolation schemas tested using three different main data, i.e. mean daily temperature, mean daily temperature at zero elevation and estimation error of T’-T, (T is mean daily temperature and T’ is estimated mean daily temperature from linear regression equation of T’=xm+b, where x is elevation). Normal data, which is mean data for mean daily temperature from at least 30 years period were also used in this study, as the deviation came from the difference of mean and normal for mean daily temperature. To generate interpolation surface for deviation of mean daily temperature from normal, interpolation schema was used twice, one for the current data and another for normal data. Cross validation analysis were used to calculate five different measures of error. The results shows that IDW interpolation technique using mean daily temperature at zero elevation was the best interpolation schema to generate spatial distribution of mean daily temperature. This schema is recommended to generate spatial distribution for deviation of mean daily temperature normal for Peninsular Malaysia.
|
|
|