Gridded Daily Rainfall Data for Ghana for the period 1960 - 2015: Approach and Validation Process
Description
Spatial interpolation of rainfall using 190 daily gauge measurements across all the four (4) agro-ecological zones of Ghana for a pertiod of 56 years (1960 -2015). It involved the estimation of a regionalised value of daily, monthly or annual at an unsampled point based on a weight of observations from station gauge data expressed mathematically with the Minimum Surface Curvature (MSC) interpolation method. A total of 190 stations across the above mentioned agro-climatic zones were utilised. The distribution of the stations over the country is depicted in Fig. 2 and the corresponding number of stations for each zone is outlined in Table 1. The record for these stations were not complete at all times as there were inherent gaps for some days and months for some stations, a challenge adequately addressed by the interpolation. A 56-year time frame (1960 -2015) of daily rainfall data for each station results in 20,496 time steps per station. Application of the Minimum Surface Curvature (MSC) interpolation at a 0.25° x 0.25° gridding resulted in a final netcdf file of daily rainfall dataset and the final output was validated using a point- pixel assessment. The Validation involved the use of 19 GMet gauge stations (10%) by comparing its Ratios, Correlations and Root Mean Square Errors to that of the gridded data to show the level of accuracy for the four (4) seasons (SON [Fall], DJF [Winter], MAM [Spring], JJA [Summer]) and the performance of the interpolation technique for each of the 19 gauge stations used for the validation.
Files
Steps to reproduce
Input data of daily rainfall data was interpolated for a 56-year period (1960-2015) over the entire country Ghana. Data homogenenisation was done using RHtestsV4 to determine stations with excessive change points which were excluded subsequently. Application of the Minimum Surface Curvature (MSC) interpolation at a 0.25° x 0.25° gridding resulted in a final netcdf file of daily rainfall dataset and the final output was validated using a point- pixel assessment. The Validation involved the use of 19 GMet gauge stations (10%) by comparing its Ratios, Correlations and Root Mean Square Errors to that of the gridded data to show the level of accuracy for the four (4) seasons (SON [Fall], DJF [Winter], MAM [Spring], JJA [Summer]).