Satellite based Oceanic Precipitation and Salinity Estimation and their Applications to Various Process Studies

Abstract

About 78% of precipitation and 86% of evaporation occur over the vast oceans which are among the major contributors of the global water cycle and influence the global climate significantly. Evaporation minus precipitation (E􀀀P), also known as the freshwater flux, is assumed to be correlated with the sea surface salinity (SSS) at large-scale which determines the density of sea water and is an important marker of the ocean circulation. Thus, accurate estimation of precipitation, evaporation and SSS is very important to infer the various aspects of the global water cycle. However, in situ observations of these parameters are meager over the large oceans and inadequate to better understand their variability and teleconnections. With the advent of earth observation satellites, it is now possible to estimate these parameters with reasonable accuracy in order to better understand their spatial and temporal variability at global and regional scales. In the first part of the thesis, precipitation is estimated over the Indian monsoon region using the first dedicated Indian meteorological geostationary satellite 􀀀 Kalpana-1. Two algorithms are used to estimate precipitation using this satellite data. First one is based on the state-of-the-art method 􀀀 GOES (Geostationary Operational Environmental Satellite) Precipitation Index (GPI) which uses thermal infrared channel of Kalpana-1 satellite for precipitation estimation at 1 latitude 1 longitude resolution. This Kalpana-1 derived precipitation product is compared with the multisatellite precipitation product from the Global Precipitation Climatology Project (GPCP) and available rain gauge data for the southwest monsoon season of 2009 at various temporal scales such as weekly, monthly and seasonal. Results show that this precipitation product picks up active and break spells of monsoon very well and performs reasonably well for the precipitation estimation at larger spatial and temporal scales. In order to estimate precipitation at finer spatial and temporal resolutions, a new algorithm namely INSAT (Indian National Satellite System) Multispectral Rainfall Algorithm (IMSRA) is developed which benefits from the high spatial and temporal resolutions of Kalpana-1 satellite and more accurate precipitation estimates from the Tropical Rainfall Measuring v Mission (TRMM) 􀀀 Precipitation Radar (PR) data. This algorithm also uses proper cloud classification schemes from thermal infrared and water vapour channels of Kalpana-1 and is able to estimate precipitation at high spatial (0.25 latitude 0.25 longitude) and temporal (3􀀀hourly) resolutions. The estimated precipitation from this algorithm is able to reproduce the detailed synoptic features of the Indian summer monsoon reasonably. The estimated precipitation from this new algorithm is evaluated with independently developed global multisatellite precipitation products viz., TRMM Multisatellite Precipitation Analysis (TMPA) and GPCP, and rain gauge data for two southwest monsoon seasons of 2008 and 2009 at daily and monthly scales. Results indicate that the IMSRA performs reasonably well over non-orographic regions and has potential for small-scale precipitation estimation. But, this algorithm substantially underestimates precipitation over orographic regions like along the west coast of India and foothills of the Himalayan regions which is inherent problem with infrared based precipitation estimation techniques. Furthermore, a number of global satellite-based precipitation products are available after the launch of the TRMM satellite. But to find the widest usage and applications of these precipitation products, their extensive evaluation is needed. In this direction, three TRMM-based precipitation products namely, TRMM Microwave Imager (TMI)-derived 2A12, TMPA version 6 and Global Satellite Mapping of Precipitation (GSMaP) are compared with the first space-based active microwave PR-derived 2A25 precipitation over the north Indian Ocean under the cyclonic conditions. Results show that all the three precipitation products systematically underestimate precipitation as compared to the PR. However, the TMPA shows good agreement with the PR during low to moderate precipitation regimes whereas the GSMaP is in better agreement with the PR than the TMPA during the extreme precipitation events. Moreover, the TRMM satellite provides a unique opportunity to estimate evaporation using a single radiometer 􀀀 TMI. Evaporation is estimated from the TMI measurements at monthly scale and at 0.25 latitude 0.25 longitude resolution over the tropical oceans using bulk aerodynamic formula for the period 1998-2010. The estimated vi evaporation and TMPA-3B43 precipitation are used to estimate freshwater flux over the tropical oceans for the study period. The estimated evaporation, precipitation and freshwater flux are compared with the independently developed satellite-based Hamburg Ocean Atmosphere Parameters and Fluxes from Satellite Data (HOAPS􀀀3) product and also validated with in situ observations from the moored buoys. The relationship between local E􀀀P and SSS is also investigated quantitatively over the north Indian Ocean which shows a considerable linear relation between them although it exhibits notable regional and seasonal variations. Moreover, recent studies suggest that the outgoing longwave radiation (OLR) can be used for the SSS estimation. Hence, a study to estimate SSS in the Bay of Bengal during the southwest monsoon by the combined use of E􀀀P and OLR is carried out. The comparison of the estimated SSS with in situ observations reveals that the SSS can be estimated at large-scale from the combined use of E􀀀P and OLR with reasonable accuracy. However, the SSS at finer spatial and temporal scales can only be achieved by the earth observation satellites. With the launch of the first salinity satellite mission Soil Moisture and Ocean Salinity (SMOS) and the Aquarius mission thereafter, it is now possible to produce comprehensive global maps of SSS. However, the validation of the SMOS-derived SSS over the Indian Ocean showed rather higher error over the north Indian Ocean than expected. Hence, an attempt has been made to estimate SSS over the tropical Indian Ocean (TIO) by the synergistic use of in situ observations from the buoy data and SMOS-derived SSS through objective analysis. A preliminary analysis is done in the TIO at monthly time scale for the year 2010 at 0.25 latitude 0.25 longitude resolution. Results indicate that the analyzed SSS estimate takes the advantages of high spatial coverage by the satellite and accurate measurements of buoys and has potential for finally a better SSS estimate. In addition to the freshwater flux and SSS, sea surface temperature and sea level anomaly (SLA) are also considered as the climate change indicators. SLA plays a critical role for changes in the global climate system and is one of the important indicators of the internal adjustment of the ocean mass through the variation of temperature and salinity. The temporal evolution of the SLA is determined by the combined e ect of eustatic vii (freshwater flux), halosteric (salinity), thermosteric (temperature) and divergence. In the second part of the thesis, the changes in these components are examined over the TIO from the multisatellite measurements during the contrasting Indian summer monsoon years and positive/negative Indian Ocean dipole (IOD) years. Interestingly, the results show that these parameters exhibit opposite behavior during the contrasting Indian summer monsoon and IOD years. Furthermore, recent studies suggest that the southwest and northeast monsoon rainfall over India shows long-term changes under the current global warming scenario. However, the long-term changes in the southwest and northeast monsoon rainfall over the Indian Ocean have not been investigated so far due to paucity of in situ data. Merged gauge-satellite datasets, available since 1979, has the greatest potential for precipitation monitoring at di erent time scales which enhances the relative benefits and reduces the disadvantages of each source type. Hence, it is more relevant to examine the long-term changes of precipitation over the TIO and surrounding land regions using the available multisatellite precipitation products. The long-term changes in the southwest and northeast monsoon rainfall over the TIO are also investigated from the merged satellite-based rainfall analysis from the period 1979-2010 which shows a decreasing trend in the southwest monsoon rainfall along the west coast of India and adjoining eastern Arabian Sea. However, the northeast monsoon rainfall shows a significant increasing trend in rainfall rate of about 0.5 mm day􀀀1 decade􀀀1 over a large region bounded by 10 S􀀀10 N and 55 E􀀀100 E during the period of study. viii