The Indus River system is fed by melting of snow and glaciers and by monsoon rainfall, and both have the potential to produce catastrophic flooding. The intensity of the present flood clearly exceeded previous probabilistic estimates: river-bank protections and dam spillway were typically designed to withstand smaller discharges. This project will combine statistical analysis of rainfall, streamflow, temperature and meteorological maps with projections of Global Circulation Models to provide a reliable estimate of the typical average recurrence time of such types of events for the Indus River basin at different locations. Relevant data will be collected and organized. There is especial urgency with regard to gathering information on the flooded areas to verify the flood water levels and volumes, as well as to define (or redefine) flood risk areas. From a scientific point of view it is important to understand whether these events may become more likely with climate change.
The Indus River, one of the world?s largest irrigation networks, faces both serious scarcity of water in one season and disastrous floods in another. In late July-early Aug 2010, the worst floods in Pakistan's history killed more than 1,600 people and directly affected more than 20 million more. The destruction to the farming-industry, including harvest, and the incumbent epidemics resulted in an incommensurable socio-economic crisis for the country. Thus improved risk estimates and flood prediction at the basin scale are fundamental to achieve more sustainable and secure use of water resources, and to issue flood warning as well as to guide proper design of protection and hydraulic works.
Improved understanding of complex river flow dynamics leading to such disasters will also be instrumental for other areas of the world (in particular, in the US for the South West, California, and the Northwest), where timing of snowmelt and precipitation are key for water resources management and flood protection. From an educational viewpoint, the results of the project will be presented in first-year engineering, policy and environmental courses at Duke University. Finally, the collaboration between Pakistan and US is valuable from a social point of view because it fosters development of hydrological sciences and engineering. This project is co-funded by the Division of Earth Sciences and the Office of International Science and Engineering.
More than two-thirds of Pakistan lie in semi-arid to arid zones, with rural economies mostly dependent on summertime Monsoon precipitation and meltwater from the Himalayan upstream areas. With a total area of 1,139,814 km2 – of which 1.8% covered by glaciers – the Indus river system (Figure 1) provides much of the irrigation requirements for Pakistan (i.e., 908 mm per year versus an annual average precipitation over the basin of 423 mm). From June to September – during the Monsoon season – devastating flood events can affect the IRB due to the combined Asian Monsoonal deluges and excess meltwater from the Himalayan upstream areas of the basin. Extreme floods are historically expected to hit the Eastern tributaries of Indus, such as the Jhelum and Chenab river in the Punjab province, and the lower part of the basin in the Sindh province, while the arid Northwestern regions are usually affected by deep yet localized convective storms bringing only scarce precipitation and moderate risk of flooding (see Figure 2). During the summer of 2010 – in conjunction with the Euro-Russian blocking producing extreme temperatures over Eastern Europe – an unusual westward path of the Indian Monsoon occurred over Pakistan, where soils were already moist for an earlier cyclonic disturbance from the Arabian Sea (Figure 3). These anomalous conditions resulted in a shift of Monsoonal deluges over the arid Northwestern regions of Pakistan, causing the worst flood in the history of the Indus river basin, with more than 1,500 deaths, half a million houses destroyed, and over 18 million people affected. This project investigated whether similar shifts in extreme precipitation toward Northwestern Pakistan can become more frequent and to what extent the occurrence of flooding events over the Indus river basin can be fingerprinted from synoptic scale variables. Analyzing historical flood records in Pakistan in connection with the last 140 years variability of El Nino-Southern Oscillation (ENSO) and Indian Ocean Dipole (IOD-DMI index), we found an alternating control on the onset of floods in the IRB of moderate to-weak La Nina events and positive IOD anomalies up to 9 months before the Asian Summer Monsoon (Figure 4). Flooding events over the main Indus river basin are comparatively more rare than the ones affecting the Eastern tributaries of Indus and show an average recurrence time of about 18-30 years. We also show that starting from the 60s, IOD variability and the 18-30 years mode of the Northwestern Monsoonal precipitation have undergone significant intensification in conjunction with a weakening in the causal relationship between the Indian Monsoon and ENSO. Our results suggest that conditional probability models based on the ENSO/IOD coupled evolution can be useful to assess flood risk for the IRB and thus help prepare for disastrous flooding with sufficiently long lead time.
