Appearance of extreme monsoonal rainfall events and their impact onerosion in the HimalayaBODO BOOKHAGEN*Geography Department, UC Santa Barbara, Santa Barbara, CA 93106-4060, USA(Received 10 January 2010; in final form 15 January 2010)Monsoonal rainfall in the Himalaya dominates erosion and sediment transportthrough the fluvial system. In addition to the strong seasonal nature of the Indiansummer monsoon, striking interannual variations in monsoonal strengthcharacterize longer records. For example, during any given year, rain maypenetrate further into the orogen, even though peak rainfall amounts almostalways occur at topographic barriers in regions with high relief, regardless ofoverall monsoonal strength. Tropical Rainfall Measurement Mission (TRMM)product 3B42 rainfall is first used to document the spatial rainfall distribution andthen the TRMM time series are used to identify temporal and seasonal rainfallvariations. A simple, but robust magnitude-frequency relation for each rainfallpixel is used to show that rainfall greater or equal to the 90th percentile occurs atleast twice as often in mountainous terrain as in low-elevation regions. Previousfield-based observations and measurements show that this significantly highernumber of extreme events leads to higher erosion volumes and greater fluvial-mass transport rates. The spatiotemporal context of these extreme events helps topredict occurrences of high sediment flux and could underpin the strategicdevelopment of preventative measures. Improved statistics for extreme events arekey to mitigating the filling of hydropower reservoirs and abrasion of hydropowerturbines, as well as to sustaining infrastructure and successful agriculture in thedownstream section of the Himalaya.1. IntroductionShort-lived extreme weather events exert control on mass-transport processes and,hence, profoundly impact the character and rates of surface erosion processes (e.g.Baker and Kale 1998, Coppus and Imeson 2002, Hartshorn et al. 2002, Dadson et al.2003, Snyder et al. 2003, Bookhagen et al. 2005a, Lague et al. 2005). In theHimalaya, interannu al variations in the strength of the Indian summer monsoonstrongly influence sediment flux and river discharge to the foreland (e.g. Sah andMazari 1998, Paul et al. 2000, Barnard et al. 2001, Gabet et al. 2004, Bookhagenet al. 2005a, Singh et al. 2007). During large floods, as well as during abnormally wetmonsoon years with elevated rainfall, large sediment volumes are transportedthrough the fluvial system and transiently stored in the low-gradient reaches beforeultimately reaching the oceans (e.g. Goodbred 2003). The rate of sediment transport*Email: [email protected], Natural Hazards and RiskVol. 1, No. 1, March 2010, 37–50Geomatics, Natural Hazards and RiskISSN 1947-5705 print/ISSN 1947-5713 online ª 2010 Taylor & Francishttp://www.tandf.co.uk/journalsDOI: 10.1080/19475701003625737Downloaded By: [Bookhagen, Bodo] At: 23:17 25 March 2010within the hydrologic cycle impacts both a diverse ecosystem and a population ofmore than 1 billion people (Ives and Messerli 1989, Stern 2007).Flooding in the mountainous Himalaya and adjacent low-relief areas can beattributed mainly to two causes: (1) heavy or extreme rainfall events associated withsynoptic climatic patterns; and (2) artificial and natural dam bursts. Both processeslead to a sudden increase of water overwhelming the fluvial system and increasing thefluvial transport capacity. Flood prediction related to extreme rainfall events isdifficult because forecasting abilities of monsoonal rainfall are limited. Although theaberrant behaviour of the Indian monsoon is well documented, its characteristicsand nature are not yet fully understood (e.g. Charles et al. 1997, Krishnamurthy andShukla 2000, Gadgil et al. 2003, Kulkarni et al. 2009, Singh et al. 2009, Wulf et al. inpress). Because both local factors and global teleconnecti ons influence monsoonstrength, a successful prediction of the Indian monso on season is commonly deemedto be almost impossibl e (e.g. Webster 1987, Webster et al. 1998, Francis and Gadgil2009, Rahman et al. 2009). In particular, rainfall observation and prediction inremote, high mountain terrains need to be improved. Coupled with extreme clim aticevents are mass-transpo rt events, such as those triggered by excess rainfall amounts.A bette r assessment of the connection between these processes is desirable, but oftendifficult to achieve due to the lack of adequate monitoring possibilities.The second flooding cause is catastrophic dam bursts often associated with glaciallake outburst floods (GLOFs) in the Himalaya (e.g. Richardson and Reynolds 2000,Kattelmann 2003, Meyer et al. 2006, Bajracharya et al. 2007). However, landslidelake outburst floods (LLOFs) are equally important and have a similar destructivenature (e.g. Shang et al. 2003, Dunning et al. 2006, Gupta and Sah 2008). In manycases, the formation of a lake and its rate of growth in remot e, mountainous regionscan be observed with repeated satellite imagery (e.g. Ka¨a¨b and Vollmer 2000, Ka¨a¨b2002, Quincey et al. 2005, Bolch et al. 2008, Leprince et al. 2008, Scherler et al. 2008).If appropriate observations and warning mechanisms a re in place, disasters can beavoided (e.g. Richardson and Reynolds 2000, Stern 2007, Korup and Clague 2009,Stone 2009). Today, post-flood identification is possible with multispectral satellite-or aerial-image analysis in the Himalaya. However, relating erosion processes withclimatic triggering factors and developing a quantitative unde rstanding of systemresponse (the Earth’s surface) to forcing magnitudes (rainfall amounts) requiresdifferent datasets and approaches.In this study, rainfall data are utilized from the Tropical Rainfall MeasurementMission (TRMM) satellite, a joint mission between the US National Aeronauticsand Space Administration (NASA) and the Japanese Aerospace Exploration Agency(JAXA). These data, recorded at 3-h intervals allow the identification of extremeevents and their spatiotemporal context. Whereas these data cover onl y the last 11years (1998–2008), several distinctive rainfall patterns help to delineate generalrelationships between rainfall, topographic relief and extreme-rainfall events. Withthis analysis, extreme events are not attempted to be predicted, but rather provide aframework for likely times and locations of extreme events.This study falls in line with recently emerging