Abstract
The problem of obtaining the area average rainfall from scattered raingauge data on to a regular grid has been addressed. A new interpolation formula has been obtained using finite element method giving due consideration to missing observations in any raingauge station data time series. An important feature of this method is that there is no need to redesign the triangulation of the given region in case: (a) if number of observations increase in this region, or (b) if observations are missing in the data time series of any station chosen as a node in triangulating this region. Their contribution to the areal precipitation has been incorporated in a mathematically consistent manner. As an application of this method, daily rainfall time series of 119 raingauge stations for four successive summer monsoons (1986–1989) are utilized to obtain the daily precipitation values on a spatial grid with 1.8° × 3.6° (lat/lon) resolution. From the daily gridded rainfall, pentads (5-day averages) of rainfall are obtained. These pentads are arranged in a time series and a principal component analysis (PCA) is performed. The first four principal components explain 63% of the total variance of rainfall over India. The oscillations depicted by these modes agree well with earlier studies. It is concluded that a few dominant PCs of rainfall calculated from pentads (as compared to seasonal and annual average rainfall) explain a larger part of the total variance of rainfall. This method is essentially related to large-scale rainfall depiction. It could, therefore, easily be used to produce global gridded precipitation fields from raingauge measurements. The spatial interpolation code is available on request from the first author.
Get full access to this article
View all access options for this article.