H13D-0950
Trends in mean and extreme rainfall in South Florida and their correlations with sea surface temperature anomalies
Historical time series of daily rainfall in South Florida over the last half century were used to extract trends in mean and extreme rainfall, specifically, the trends in moving means, moving standard deviations and moving return levels. The Generalized Pareto Distribution developed within extreme value theory was used to compute the return levels within each moving window. The influence of threshold choice and missing value filling methods on the return levels were studied. Extreme value theory was applied to the entire time series after removing seasonality, while both mean and extreme trends were calculated by using wet and dry season data separately. The relation of mean and extreme rainfall in South Florida with sea surface temperature anomalies were investigated by computing correlations of a variety of indices corresponding to the latter with the moving mean, standard deviation and return levels of rainfall. New insights were obtained about the mean and extreme rainfall trends as well as their correlations with Atlantic and Pacific surface temperature anomalies.
H13D-0951
Runoff Modeling With Flow Dependent Response Functions
Reliable prediction of peak flows is essential for safe and efficient management of water resources. Calibrations of hydrological models based on a range of stream flows indicate that the volumetric error of the predicted size of the spring flood in Sweden can be as large as 20%. A significant part of this error originates from simplifications in the spatial and hydrodynamic description of watercourse networks, as well as statistical problems to give proper weight to extreme flows. This study formulates a methodology to make compartmental models more reliable for peak flows with support in hydrodynamic theory. A 1-D runoff model for the sub-catchment is parameterized in a form appropriate for a compartmental model by analyzing the distribution of residence times within the watercourse network. This distribution depends on the topologic characteristics of the watercourse network as well as flow regime stage. Thus, the particular geomorphological characteristics of the watercourse network and how these properties vary with flow regime are incorporated into the response functions of the compartmental model. A routing routine was used to numerically investigate how the response time characteristics in seven subcatchments in Rönne River basin, Sweden, varied with stage. Simplified hydraulic analysis was employed for the parameterization. The derived model is compared with an identical compartmental model with a response function that is independent of water level. The results show that the previous model version provides the best model behavior during high flows, since it accounts for stage. This is particular important during extreme flows when a significant portion of the water flows outside the normal stream channels. The results call attention to the importance of accounting for the hydraulic-hydromorphological properties of watercourse networks to obtain correct response times, especially so in more complicated stream networks.
H13D-0952
Use of Generalized Extreme Value Covariates to Improve Estimation of Trends and Return Frequencies for Lake Levels
One of the most important tools in water management is the accurate forecast of both long-term and short- term extreme values for both flood and drought conditions. Traditional methods of trend detection, such as ordinary least squares (OLS) or the Mann-Kendall test, are not aptly suited for hydrologic systems while traditional methods of predicting extreme flood and drought frequencies, such as distribution fitting without parameter covariates, may be highly inaccurate in lake-type systems, especially in the short-term. In the case of lakes, traditional frequency return estimates assume extremes are independent of trend or starting lake stages. However, due to the significant autocorrelation of lake levels, the initial stage can have a significant influence on the severity of a given event. The aim of this research was to accurately identify the direction and magnitude of trends in flood and drought stages and provide more accurate predictions of both long-term and short-term flood and drought stage return frequencies utilizing the generalized extreme value distribution with time and starting stage covariates. All of the lakes researched evidenced either no trend or very small trends unlikely to significantly alter prediction of future flood or drought return levels. However, for all of the lakes significant improvement in the prediction of extremes was obtained with the inclusion of starting lake stage as a covariate. Traditional methods of predicting flood or drought stages significantly overpredict stages when starting lake stages are low and underpredict stages when starting stages are high. The difference between these predictions can be nearly two meters, a significant amount in urbanized watersheds in areas of the world with flat topography. Differences of near two meters can mean significant alterations in evacuation or other water management decisions. In addition to improving prediction of extreme events, utilizing GEV with time or starting stage covariates can provide guidance in lake management decisions in regards to how much water to release from a lake in preparation for an approaching hurricane, appropriate lake levels to maintain throughout the year or determining minimum structure floor elevations in the watershed and allow more accurate forecasting of future water supply or impacts to tourism. The methods utilized in this research to determine lake level return period of flood and drought can be applied to nearly any region globally.
H13D-0953
Simulation of 2D Fields of Raindrop Size Distributions
The raindrop size distribution (DSD hereafter) is of primary importance for quantitative applications of weather radar measurements. The radar reflectivity~Z (directly measured by radar) is related to the power backscattered by the ensemble of hydrometeors within the radar sampling volume. However, the rain rate~R (the flux of water to the surface) is the variable of interest for many applications (hydrology, weather forecasting, air traffic for example). Usually, radar reflectivity is converted into rain rate using a power law such as Z=aRb. The coefficients a and b of the Z-R relationship depend on the DSD. The variability of the DSD in space and time has to be taken into account to improve radar rain rate estimates. Therefore, the ability to generate a large number of 2D fields of DSD which are statistically homogeneous provides a very useful simulation framework that nicely complements experimental approaches based on DSD data, in order to investigate radar beam propagation through rain as well as radar retrieval techniques. The proposed approach is based on geostatistics for structural analysis and stochastic simulation. First, the DSD is assumed to follow a gamma distribution. Hence a 2D field of DSDs can be adequately described as a 2D field of a multivariate random function consisting of the three DSD parameters. Such fields are simulated by combining a Gaussian anamorphosis and a multivariate Gaussian random field simulation algorithm. Using the (cross-)variogram models fitted on data guaranties that the spatial structure of the simulated fields is consistent with the observed one. To assess its validity, the proposed method is applied to data collected during intense Mediterranean rainfall. As only time series are available, Taylor's hypothesis is assumed to convert time series in 1D range profile. Moreover, DSD fields are assumed to be isotropic so that the 1D structure can be used to simulate 2D fields. A large number of 2D fields of DSD parameters are generated and the corresponding reflectivity fields are derived. They are in good agreement in terms of mean, standard deviation and variogram with reflectivity fields measured by an operational radar during the same event. This shows the promising potential of the proposed DSD-field simulation approach.
H13D-0954
Extreme Storm Event Assessments for Nuclear Facilities and Dam Safety
Extreme storm events over the last 35 years are being assessed to evaluate flood estimates for safety assessments of dams, nuclear power plants, and other high-hazard structures in the U.S. The current storm rainfall design standard for evaluating the flood potential at dams and non-coastal nuclear power plants is the Probable Maximum Precipitation (PMP). PMP methods and estimates are published in the National Weather Service generalized hydrometeorological reports (HMRs). A new Federal Interagency cooperative effort is reviewing hydrometeorologic data from large storms which have occurred in the last 20 to 40 years and were not included in the database used in the development of many of the HMRs. Extreme storm data, such as the January 1996 storm in Pennsylvania, June 2008 Iowa storms, and Hurricanes Andrew (1992), Floyd (1999), Isabel (2003), Katrina (2005), need to be systematically assembled and analyzed for use in these regional extreme storm studies. Storm maximization, transposition, envelopment, and depth-area duration procedures will incorporate recent advances in hydrometeorology, including radar precipitation data and stochastic storm techniques. We describe new cooperative efforts to develop a database of extreme storms and to examine the potential impacts of recent extreme storms on PMP estimates. These efforts will be coordinated with Federal agencies, universities, and the private sector through an Extreme Storm Events Work Group under the Federal Subcommittee on Hydrology. This work group is chartered to coordinate studies and develop databases for reviewing and improving methodologies and data collection techniques used to estimate design precipitation up to and including the PMP. The initial effort focuses on collecting and reviewing extreme storm event data in the Southeastern U.S. that have occurred since Tropical Storm Agnes (1972). Uncertainties and exceedance probability estimates of PMP are being explored. Potential effects of climate variability and change on the PMP are also under investigation.
H13D-0955
Probabilistic Rainfall Simulation: Structural Analysis for Crop Index Insurance
A new probabilistic methodology for rainfall simulation/record extension is presented here. The methodology extends the rainfall simulation component of the NOAA Ensemble Streamflow Prediction system which operates using the historical record of precipitation and temperature in combination with the current conditions to produce an ensemble of precipitation time series. In this methodology, a spatial relationship among adjacent stations with longer records is embedded to further enhance the result. Cross-ensemble- member probabilities are updated given spatial proximities. The method could be used to generate rainfall simulations whereby rainfall occurrence and amounts are modeled. The methodology has been applied in Adi Ha in Northern Ethiopia to extend the rainfall record. The extended record will be used to design a weather insurance contract for farmers in the area.
H13D-0956
Climate Dynamics of Regional US Southeastern drought
The phenomena of droughts both regional and continental have received considerable attention from both science and policy. Understanding the larger scale dynamics of these events is critical to improving predictability, management and mitigation strategies. The history of drought in the United States shows both long (multi-decadal) and short (seasonal or yearly) droughts in various regions. Some of the most severe droughts, such as those with the largest economic losses or that have received the most attention both from the scientific and broader communities have occurred in the Midwest/Great Plains (e.g. the "Dust Bowl" years, the 1988 drought) and generally correlate with continent-wide anomalies. The Southeast region of the US, while generally having a more humid temperature climate than the rest of the country, also is subject to periods of drought conditions. In this study we analyze long term records of PDSI in the southeastern United States. While some occurrences correlate with larger continental scale droughts, many severe southeastern droughts occur during a synoptic pattern correlating with wetter conditions through the greater Midwest, have a seasonal pattern different than larger continental scale anomalies, and show correlations with patterns in Atlantic tropical activity. While winter and spring deficits may initiate a drought, we find a proportionally larger decrease in summertime precipitation during severe drought periods. Some of this decrease may be related to the contribution of tropical systems, which increases in the periods following droughts. We suggest that the dynamics of drought in this region differ from the larger US pattern, and particularly need to account for the interaction between continental and tropical contributions. With an increasing population and areas of high agricultural productivity, we argue this region deserves further attention from both the scientific and larger community, that understanding these dynamics will become increasingly important and that the observed patterns can be applied to create more successful drought mitigation and onset forecasts.
H13D-0957
Impacts of Mountain Pine Beetle on Peak Flow in the Fraser Basin in British Columbia
Increasing winter temperature in combination with forest management practices are the main drivers for the Mountain Pine Beetle (MPB) epidemic in British Columbia (BC). The infestation of MPB has now turned into a major threat to the natural habitat of the province. The Fraser basin, the largest watershed in BC, is the most affected watershed with an infested forest area of 7.7 million hectares (88% of the watershed) [Redding and Pike 2007]. Forest cover is a key modifier of the watershed's peak flow regime. The peak flow generally increases when forest cover is reduced. Major parts of the Fraser basin have only a limited number of gauging stations (or are even ungauged). The goal of the project was to develop a hydrological model that can predict peak flow increases but does not rely on complex data inputs for its validation and calibration procedures. The model consists of two major components: climate input and runoff. The climate input component determines the mean annual snowmelt as well as the maximum rainfall based on long term climatic averages. This information is then used to determine the time and the capacity of the peak flow for every 3rd order watershed. The runoff component delineates hydrologic processes such as Hortonain Overland Flow, Saturation Overland Flow and Shallow Surface Flow. The model combines the two components and computes a map of peak flow contribution. A peak flow analysis has been carried out to validate the model results using available gauging stations in subcatchments. The validated model has been then applied to the entire watershed to analyze the impacts of MPB on peak flow in the Fraser basin. The presentation will show the conceptual presentation of the hydrological model. It will highlight the results of the peak flow analysis and show initial results of the application of the model. Cited Literature: Redding, T. and Pike, R (2007). Mountain Pine Beetle and Watershed Hydrology Workshop Summary, Streamline Watershed Management Bulletin 11(1), pp.22-24.
H13D-0958
Effect of Space-Time Rainfall Resolution on the Prediction of Floods – A Simulation Case Study
Satellite-borne rainfall estimation, together with land surface datasets, allows for the development of a quasi- global flood monitoring and forecasting systems. When combined with a parsimonious hydrological model, they have the potential of providing prediction of extreme flood events even for ungauged basins. This potential lends the question of how accurate such forecasting systems can be. The use of such data in an operational flood forecast system requires a better understanding of the effect of data uncertainty on the prediction of peak discharge and times to peak for different basin scales. In order to investigate the effect of temporal and spatial aggregation of rainfall fields we use NEXRAD rainfall data to force the link-hillslope scale hydrologic model implemented as CUENCAS for the 1100 km2 Whitewater basin. This area was selected by the Hydro-Kansas research group to explore the physical basis of statistical self-similarity in peak flows and cover a sufficient range of spatial scales for this study. Radar rainfall information was considered error-free and aggregated to generate new rainfall maps with lower spatial (Äs) and temporal (Ät) resolution. Flow predictions were generated for different basin scales and rainfall map resolutions. The resolution analyzed ranges from Äs= 1 arcmin and Ät=15min to Äs= 15 arcmin and Ät=3h, that correspond to the radar and satellite product resolution, respectively. We focus our analysis on the prediction of peak flows and times to peak for a large range of basin areas, as a way to identify the smaller scale for which remotely sensed rainfall information can provide accurate prediction of the hydrological basin response. Our results indicate that the temporal and spatial resolution of the rainfall data strongly shape the hydrological response of river basins. It also shows that this control depends strongly on basin size. Our study reveals the limitations of using coarse resolution forcing data in the accuracy of predicted peak discharges and times to peak. It provides initial clues about the range of scales of applicability of satellite-borne rainfall estimations.
H13D-0959
Monitoring, Analyzing and Modeling Flash Flooding Processes of a Small Mountainous Catchment in Hong Kong
Due to frequent heavy monsoon rainfalls in the rainy seasons and the steep topographic slopes, Hong Kong suffers the hazards of flash floods almost every year. This study focuses on the understanding of the features of flash flooding events occurred in a small mountainous catchment, with an area of 0.08 km2 and an average slope of 35.1 °, in Hong Kong. In Mid-2007, two automatic rain gauges, one in open air and one under a tree, and one automatic water level data-logger were installed at this small catchment for providing one-minute time step data. Since then, a number of severe rainfall hyetographs and related runoff hydrographs have been obtained. In these observations, the maximum rainfall intensities are 4.5 mm/min for a one-minute step, 3.4 mm/min for a five-minute step, and 1.2 mm/min for a one-hour step. Accordingly, the water level can rise sharply to peak within 15 min. By applying a baseflow separation technique to those storm hydrographs, in the study catchment, it is found that the ratios of the direct runoff to the corresponding total rainfall volume vary around 2%, which can further be functioned by two factors, the total rainfall volume and the antecedent baseflow. Moreover, a rainfall-runoff model, TOPMODEL, is applied to simulate those flash flooding processes. This study shows that the model is able to simulate the quick response of the relationships of rainfall and runoff at a one-minute time step. Through revisiting the representation of the physical processes of TOPMODEL, we will also examine the pros and cons of the model in simulating flash flooding processes at such a fine time step, and the further results will be presented in the conference.
H13D-0960
Precipitation Extremes in Washington State: Are they changing?
Estimation of precipitation extremes is critical to the design of stormwater infrastructure. The standard approach is to estimate frequency distributions of precipitation extremes (e.g., annual maximum or peaks over threshold) for given accumulation intervals (1, 2, 3, 6, …. hours), and to design stormwater structures to withstand the critical duration for a given return period. An alternative, more sophisticated approach, is to simulate stormwater system performance for a sequence (time series) of precipitation, from which the frequency distribution of discharge is estimated. In either case, however, the frequency distribution of precipitation extremes is required, either directly or indirectly, and an underlying assumption is that the probability distribution of precipitation extremes is statistically stationary. This assumption is called into question by climate change. We examine, therefore, whether there is evidence for changes in the probability distributions of precipitation extremes for three major metropolitan areas of Washington State: Seattle, Vancouver, and Spokane, and how those distributions might change over the next half century. The historical analysis is based on hourly precipitation records for gauges surrounding the three metropolitan areas for the time period 1949-2007. Future precipitation simulations were taken from dynamically downscaled (using the Weather Research and Forecast model, WRF, implemented at roughly a 20 km spatial distribution) current and future climate output from global simulations of the ECHAM 5 General Circulation Model (GCM) for the time periods 1970-2000 and 2020-2050. Precipitation events at the 1, 2, 3, 6, 12, and 24 hour duration and the 1, 1.5, 2, 5, 10, 20, 50, 100 year return periods were analyzed. L- moment analysis of annual maximum precipitation and peak over threshold analysis were used to identify any changes in the return period and frequency of storms. Statistically significant changes in precipitation extremes were detected in both historical and modeled future precipitation events, but patterns of changes are complex, and in particular for historical precipitation, both positive and negative trends have occurred.
H13D-0961
Characterization of the Frequency of Flash Flood Occurrence through Hydrologic Modeling for Southern California
The occurrence of flash flooding is of concern in hydrologic and natural hazards science due to the top ranking of such events among natural disasters in terms of both the number of people effected globally and the proportion of individual fatalities. A better understanding of such extreme event frequency over various time and spatial scales is sought in this research as an important aspect of climate and hydrologic science. The study characterizes flash flood occurrence frequency through interdisciplinary meteorological, hydrologic, and geomorphologic modeling in the Southern California region over the historical period from 1950 to 2005. A combined modeling approach is necessitated due to (a) infrequent reporting of flash flood occurrence with high spatial detail, and (b) the relative sparseness of observed precipitation records covering regions and the spatial scales of flash flood occurrence (e.g., typically over a few tens of km2) to derive climatology based on observations only. The modeling approach includes generation of historical orographically-driven precipitation over the region with 3km spatial resolution, hydrologic modeling of soil moisture, and estimation of flash flood occurrence via a hydrologic response threshold model for small-scale watershed on the order of 30km2. The precipitation modeling is driven by NCEP Reanalysis forcing, which subsequently provides input to the hydrologic and flash flood occurrence modeling. This research represents a regional look at the spatial variability in flash flood occurrence with relatively high spatial resolution. This paper will focus on a general description of the modeling methodology and on initial results of the spatial character of flash flood occurrence frequency over the mountain-to-foothill regions of the Transverse and Peninsular mountain ranges of Southern California.
H13D-0962
Multiple Hazard Assessment of Extreme Rainfall, Winds and Flooding from Hurricane Isabel
Landfalling tropical cyclones are one of the major hazards for the eastern United States. Hurricane Isabel produced significant damage in the mid-Atlantic region in September 2003, resulting from inland flooding, storm surge flooding in the Chesapeake Bay and high winds in Virginia. Simulations of Hurricane Isabel using the Weather Research and Forecasting (WRF) model, with initial and boundary conditions generated by the Geophysical Fluid Dynamics Laboratory's (GFDL) hurricane model, are used to examine multiple hazards associated with the storm. Analyses will focus on heavy rainfall and flooding and include intercomparisons of the structure and evolution of extreme rainfall over land as simulated by WRF with analyses based on HydroNEXRAD rainfall fields and observations from rain gage networks. A particular focus of these analyses is the evolving distribution of rainfall and variation of winds, relative to the center of circulation, as the storm moves over land. Storm surge in the Chesapeake Bay will also be examined with simulations generated from the Advanced Circulation model (ADCIRC). Conclusions from these analyses will be synthesized to highlight emerging tools for assessing hazards from landfalling tropical cyclones.
H13D-0963
Using Wavelets to Evaluate Watershed Signals from Precipitation Extremes at a Localized Temporal Scale
Resource managers in the future will be required to make decisions regarding complex systems under extreme uncertainty and to evaluate the sustainability of these natural systems. The variability and extremes of precipitation will be one of the major variables impacting natural systems, and decision making. These future decisions will be evaluated based upon economic costs and benefits, and core mission valuation. This will be particularly important in evaluating the effects and impacts of climate change on natural system response. In this case study, we evaluate the signal organization and its nature within a watershed in east Tennessee. In this study, temporal analyses were conducted on weekly time series data of water chemistry (nitrate, chloride, sulfate and calcium concentrations) collected from November 1995 to December 2005 at the West Fork of Walker Branch in Oak Ridge, Tennessee (Mulholland 1993, 2004). Hydrochemistry plays an important role in ecosystem services, particularly nitrate (Mulholland et al. 2008), and in general the signal responses can be complex. The time series in this study was modeled using a wavelet approach as a mechanism for evaluating short-term temporal effects. In general, time series signals of watershed hydrochemistry may provide clues as to broad environmental, ecological and economic impacts at the basin scale. References: Mulholland, P.J. (1993), Hydrometric and stream chemistry evidence of three storm flowpaths in Walker Branch Watershed, Journal of Hydrology, 151: 291-316. Mulholland, P.J. (2004). The importance of in-stream uptake for regulating stream concentrations and outputs of N and P from a forested watershed: evidence from long-term chemistry records for Walker Branch Watershed, Biogeochemistr. 70: 403-426. Mulholland, P.J., A.M. Helton, G.C. Poole, R.O. Hall Jr., S.K. Hamilton, B.J. Peterson, J.L. Tank, L.R. Ashkenas, L.W. Cooper, C.N. Dahm, W.K. Dodds, S.E.E. Findlay, S.V. Gregory, N.B. Grimm, S.L. Johnson, W.H. McDowell, J.L. Meyer, H.M. Valett, J.R. Webster, C.P. Arango, J.J. Beaulieu, M.J. Bernot, A.J. Burgin, C.L Crenshaw, L.T. Johnson, B.R. Niederlehner, J.M. O'Brien, J.D. Potter, R.W. Sheibley, D.J. Sobota, and S.M. Thomas (2008). Stream denitrification across biomes and its response to anthropogenic nitrate loading, Nature, 452(13): 202-206.
H13D-0964
Influence of Sample Volume on Disdrometer and Radar Reflectivity Estimates
Commonly used surface disdrometers count the number and size of raindrops passing a plane or impacting a surface. These discrete drop measuring instruments have sampling volumes of the order 1 cubic meter when dwelling for 1 minute and observe about 1000 raindrops when the rain rate is 10 mm/hr. One difficulty in relating discrete measuring instruments to constant volume instruments, such as radars, is that the sample volume with surface disdrometers is diameter dependent with larger raindrops having larger sampling volumes because they travel further distances during fixed observational dwell times. It is well documented that disdrometers underestimate the number of larger raindrops in naturally occurring rainfall because of their small sampling volumes. This underestimation leads to biased integrated moments and larger uncertainties when a limited number of raindrops are sampled. Vertically pointing radars (VPR) sample a constant volume of the order 10,000 cubic meters at a range of 300 meters, observing about 1 million raindrops in the radar pulse volume when the rain rate is 10 mm/hr. These larger sample volumes with more sampled raindrops provide unbiased estimates of stationary raindrop distributions after converting the measured Doppler velocity power spectrum into raindrop distributions. The biases and uncertainty in the disdrometer observations are due to the uncertainty of the drop counts in each diameter size channel. The uncertainties in the VPR retrieved raindrop distribution are due to power uncertainties in each Doppler velocity channel. Before these discrete and volume-distributed raindrop phenomenon uncertainties can be explored in detail, the biases and uncertainties of small sample volume disdrometers must be understood. Therefore, this study developed a Monte Carlo Simulation that incorporates the diameter dependent sample volumes of a Joss-Waldvogel Disdrometer (JWD) and uses Poisson sampling statistics to estimate reflectivity biases and uncertainties for the JWD. The simulations and simultaneous radar/JWD observations show biases and uncertainties decreasing with increasing number of sampled raindrops with at least 300 raindrops needed to reduce the reflectivity bias and uncertainty below 1 and 0.8 dBZ, respectively.
H13D-0965
Modelling and managing runoff processes in peri-urban area
Nowadays, a deeper knowledge of the extreme runoff generation requires more inclusive and interactive understanding of its numerous determining factors. This includes not only a better estimation of meteorological extremes under changing climate conditions, but also a better evaluation of infiltration and saturation excesses, of subsurface return flows, as well as, of human impacts on surface runoff. This communication presents a physically based and spatially distributed numerical model for simulation of the hydrologic interactions between the surface and subsurface flows. Further particularities of this model correspond to: (1) a new methodology for the estimation of the precipitation input; and (2) a new modelling methodology to design Sustainable Urban Drainage Systems (SUDS) in urban and peri-urban areas. The multifractal frequency analyses have been used to evaluate the maximum precipitation rate for several durations with the design return period. This method has the advantage to rely on a few robust exponents that are physically meaningful and can be evaluated on discontinuous and/or low frequency samples. The design of Sustainable Urban Drainage Systems (SUDS) in urban and peri-urban areas with low permeability soils as well as with high groundwater levels can be used to decrease the floods risk in the inundated zones. Our model was particularly oriented towards the retention in ponds and swales, infiltration into the ground and drainage through perforated pipes to manage the storm water runoff. The methodology explicitly takes into account the interactions with the water table, the evolution of the latter with infiltration and the soil profile. Using GIS, we visualise the resulting runoff processes together with the evolution of water table levels for the two case studies: a county contiguous to Paris (France) and in the Panola Area (USA). The obtained results demonstrate the effectiveness of SUDS in urban and peri-urban areas and fluvial retention measures to attenuate floods in small urban catchments. Comparisons with natural catchments with low urban development illustrate the impact of climate change and urbanisation on extreme runoff characteristics.
H13D-0966
Exploring the Interplay Between Successive Storms and Basin Drainage Topology During the 2008 Extreme Flooding Conditions in Eastern Iowa
Historical and real-time data collected during the 2008 floods in Eastern Iowa show that small basins in the region experienced mild flooding conditions with small return periods (< 50 years), while large basins experienced extraordinary flooding conditions (> 100 years). The individual storms that preceded the flood were not extraordinarily large; rather, the week prior to the flood was characterized by a series of storm systems that successively hit the region. This observation suggests the hypothesis that extreme flooding conditions in locations draining large basins can be created by consecutive "ordinary" storms rather than by extraordinarily large storms that have been the focus of previous studies. Current analyses linking storm return periods to flood return periods do not account for the possibility that large floods can be created by consecutive storms. We perform a series of numerical simulations that show that the rainfall totals observed in June do not account for the extreme flooding that was observed, instead the timing of the individual events was compounded by the topology of the river network draining the landscape. This result highlights the need for a more precise understanding of the changes in rainfall patterns in the presence of climate change. Our simulations indicate that large basins are more sensitive to how rainfall falls over the basin rather than to the total amount of rainfall.