Akademik Veri Yönetim Sistemi
JOURNAL OF APPLIED METEOROLOGY, cilt.36, sa.12, ss.1712-1724, 1997 (SCI-Expanded)
A cumulative semivariogram (CSV) method is proposed as an optimum analysis technique for producing gridded fields of meteorological regional variables that are sampled at irregular sites as sparse data. After having discussed the basis of the CSV and its theoretical representations by functional models, the procedure of obtaining weighting functions useful for simple optimum analysis calculations from the CSVs is explained. The experimental CSVs are obtained from monthly rainfall data for northwestern Turkey. Following the interpretation of these experimental CSVs, they are converted into experimental weighting functions necessary for optimum analysis. Comparison of these experimental functions is made on an individual monthly basis with other mathematically simple but geometric weighting functions that are available in the meteorology literature. It is observed that none of the available geometric weighting functions represents completely the regional variation within one month. However. the experimental CSV weighting functions represent regional variability and remain within the domain of various available geometric models. Finally, the rainfall contour maps are produced by using the experimental CSV weighting functions for each month for northwestern Turkey.