Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Royaume-Uni → 2017 → Simulating river runoff and terrestrial water storage variability in data-scarce semi-arid catchments using remote sensing

University of Leicester (2017)

Simulating river runoff and terrestrial water storage variability in data-scarce semi-arid catchments using remote sensing

Najmaddin, Peshawa Mustafa

Titre : Simulating river runoff and terrestrial water storage variability in data-scarce semi-arid catchments using remote sensing

Auteur : Najmaddin, Peshawa Mustafa

Université de soutenance : University of Leicester

Grade : Doctor of Philosophy (PhD) 2017

Résumé
Remotely sensed data can be used as an alternative to ground based observations to predict river discharge and water storage variability. The latter dataset used consists of meteorological records from four stations (2003-2014) and daily river discharge records from one stations (2010-2014). A model was developed named ‘Leicester Model for Semi-Arid Region’ (LEMSAR). It was applied in the semi-arid Kurdistan region of Northern Iraq. TRMM Multi-satellite Precipitation Analysis (TMPA) data products (TMPA 3B42 and 3B42RT) were used with and without a bias correction. The uncorrected TMPA underestimated observed mean catchment rainfall by 10% compared to corrected data with 0.7%. Four methods of computing reference evapotranspiration (ETₒ) were applied which include Hargreaves-Samani (HS), Jensen-Haise (JH), McGuinness-Bordne(MB) and FAO Penman Monteith(PM). The variables utilised are air temperature, relative humidity and cloud cover fraction from the Atmospheric Infrared Sounder / Advanced Microwave Sounding (AIRS/AMSU), and wind speed at 10 m height from MERRA (Modern-Era Retrospective Analysis for Research and Application). Compared to ETₒ-G (PM), ETₒ-RS (HS) underestimated ETₒ-G (PM) by 3% while JH and MB overestimated by 8% to 40% at different stations. Nash-Sutcliffe Efficiency (NSE) for the LEMSAR fit with the observed hydrograph was 0.75, for a calibration period (2010-2011) using gauged rainfall data with ETₒ-G (PM). Model validation performance (2012–2014) was best (NSE =0.61) using the corrected 3B42 data with ETₒ-RS HS and poorest when driven by uncorrected 3B42RT data with ETₒ-RS JH (NSE =0.07). Data from the Gravity Recovery and Climate Experiment (GRACE : 2003-2014) were used to evaluate total water storage variability and compared with that of well observations data and LEMSAR. Trends in GRACE_TWSA were approximately -33.72 mm y-1 for the Lesser Zab catchment and -35.4 mm y-1 for the Hawler well monitoring zone while LEMSAR predicted 15 mm y-1 for the Lesser Zab Catchment. This suggest that reduction in recharge (modelled by LEMSAR) may only be responsible for about 50% of the reduction in groundwater storage. The rest could be the result of increased abstraction in response to the drought. Overall, results suggest that RS data can be usefully employed to simulate river discharge and to evaluate terrestrial water storage variability in semi-arid areas. It has the potential to help decision-makers improve water resources management

Présentation

Version intégrale (10,5 Mb)

Page publiée le 26 octobre 2018