RUNOFF ESTIMATION FOR NZOIA BASIN USING NASA’S GEOS-5 SATELLITE DATA Author: Stephen Mureithi Phone: +254 705 769 439 P.O. Box 3900-30100 Eldoret, Kenya mureithikivuti@gmail.com Abstract Never in human history has the global climate been as fragile and as precious as it is now today. With the climate fluctuating at the present era, it makes sense to constantly seek to improve our climate change resilience. This involves researching into how the weather develops climate-related risks, and taking steps aimed at coping with these risks. In Kenya, flooding is the most acute event associated with climate change resilience. Flooding occurs due to excessive runoff from a catchment area. Therefore, identifying watersheds and studying their runoff characteristics would provide valuable information to use in improving our climate change resilience. Nzoia basin exhibits temporal and spatial scarcity of rainfall and temperature data available; both for satellite and observed data. Data from rain gauge networks is insufficient at a temporal scale, and would pose challenges when used in runoff estimation. Clearly, there is a need for data with better temporal sufficiency. The use of satellite data in place of rain gauge data seems to solve the issue of temporal scarcity. National Aeronautics and Space Administration (NASA) manages the Goddard Earth Observing System model version 5 (GEOS-5) satellites. Satellites by nature are however poor on a spatial scale and therefore need area- specific calibration and validation due to the indirect nature of radiation measurements. This study aimed at improving data available from GEOS-5 satellites for use in runoff studies. By carrying out statistical analyses on temperature and rainfall data from the satellites, it is possible to calibrate this data and analyze its effectiveness in runoff estimation. With the reference being observed runoff, use of calibrated data for runoff estimation improved R-square coefficient from 0.6 to 0.9. This proves the feasibility of the study. Keywords: Climate, Runoff, Temporal, Spatial, Calibration, Big data, Small data, Radar measurement. 1 Introduction Looking down from two miles above the surface of the earth, it is impossible not to be impressed by the sheer power, grandeur and splendor of the natural world. From such a vantage point, new ways of looking at the earth, and the interconnectivity of its systems emerge. The wind becomes more potent, and it becomes easier to visualize its effect on weather patterns; and cloud formation becomes clearer to the naked eye. With such increment in information provided by a simple vantage point, it is no wonder that mountains and areas of high elevation provide the best locations for meteorological stations. Isn’t it also rather uncanny how understanding of meteorological parameters seems to be tightly linked with advances in aviation technologies? A little over 50 years ago saw the launch of the Sputnik satellite and the first human steps on the moon. For the first time, man was truly able to look back at their own planet. Soon enough radar measurement combined with computing technology gave birth to the concept of big data engineering in meteorology. Big data engineering in meteorology becomes even more necessary when the fact that the weather is affected by numerous interconnected small systems. Hypothesis free science. This is the gift big data engineering promises to offer the scientific community. This is godsend, particularly for hydrologists. In this age of satellites, data availability is no longer the limiting factor in meteorological practices. Every second satellites, by the use of radar measurement, provide additional hydrological information. With more technology being developed, this information is increasing exponentially. Datasets previously considered to be small data are increasingly becoming big data as spatial and temporal adequacy of satellites improve. Clearly, our understanding of the climate has radically changed. But that’s not all. The climate has changed too. As hydrologists and scientists rush to make sense of the rapidly growing volumes of meteorological data, one pattern has been clear for a while now; our climate resilience is rapidly being overtaken by climate fluctuations. Climate resilience defines the capacity of a socio-ecological system to absorb stresses and maintain function in the face of external stresses imposed on it by climate change. It also defines the subsequent process of adapting, reorganizing and evolving into more desirable configurations that improve the sustainability of the system, leaving