Merging satellite rainfall estimates in scarcely gauged basin : A case study of Indus basin

Rainfall is an important part of hydrological cycle and is measured normally by rain gauges, weather radars and satellite borne instruments. Ground based rain gauge data is considered to be authentic and reliable but problem arises due to its limited spatial coverage in many countries. Development of a country relies on efficient use of its water resources which is linked to the basin development. To minimize the impacts of any natural hazards such as flooding a prior knowledge of historical gauge data is required. Availability of global rainfall satellite datasets offers a wide range of ground information which is convenient these days. Bias corrected satellite datasets can provide realistic information when used in ungauged catchments. Rain gauge stations have point observations but their spatial coverage is a key limitation. This thesis focuses on using satellite rainfall products as an alternative to gauge rainfall in hydrological studies of sparsely gauged basin. Satellite products are known to behave differently at different temporal and spatial scales. The study area chosen for the research is Indus basin which is a transboundary and sparsely gauged basin. Four state of the art satellite products were used in this study namely CMORPH, PERSIANN CDR, TRMM 3B42 and MSWEP. MSWEP was used as a reference dataset for comparing the merged rainfall dataset since it is also a merged product. All these satellite products were downloaded with pixel size of 0.25° x 0.25° and daily temporal scale. Temporal scales selected were daily, weekly, fortnightly, monthly and seasonally whereas spatial scales were gauge location, zonal scales and basin scale. Indus basin was divided into four zones so that the performance of satellite and merged products can be checked. All the satellite products except MSWEP showed highest NRMSE in southern part of Indus basin which is an arid area whereas MSWEP gave high NRMSE in northern part of Indus basin which has a complex mountainous terrain. This showed that CMORPH, PERSIANN CDR and TRMM were unable to capture the true rainfall pattern in the southern part of basin. MSWEP overestimated throughout the year in the northern part. The Ratio Bias Correction method was used to correct the biasness of satellite products by correcting the products at monthly and seasonal scale. Daily NMRSE was reduced up to 20% for CMORPH, 22% for PERSIANN CDR and 14% for TRMM at Indus basin scale for monthly bias corrected data which is why monthly bias corrected data was used for merging. Merging of satellite products can be fruitful to benefit the strength of each product and minimize the weakness of products. Four different merging methods i.e. Arithmetic merging method (AMM), Inverse error squared weighting (IESW), Error variance (EV) and K Nearest Neighbor method (KNN) were used and performance was checked in two seasons i.e. non-wet and wet season. AMM and EV methods performed similarly whereas IESW performed poorly at zonal scales. The interesting find of the study KNN merging method outperformed all other merging methods and gave lowest error across the basin. Daily NRMSE was reduced to 0.3 at Indus basin scale with KNN method, AMM and EV reduced the error to 0.45 in comparison to error produced by CMORPH, PERSIANN CDR and TRMM of 0.8, 0.65 and 0.53 respectively in wet season. All merging methods gave high errors in the not wet season as compared to the wet season. Data mining tool called “Weka” was used to predict rainfall data in the validation period for K=1, 3 and 5. Predicted rainfall data gave lowest error among all merged products at zonal and basin scale. KNN merged product gave lowest error at daily scale in calibration and validation period which justifies that merging not only improves result but can also be used to predict the rainfall data in sparsely gauged basin