Search Results

You are looking at 1 - 10 of 14 items for

  • Author or Editor: Konstantine P. Georgakakos x
  • All content x
Clear All Modify Search
Konstantine P. Georgakakos

Abstract

The use of information from climate model ensemble simulations for regional hydrologic and water resources impact studies necessitates the development of diagnostic measures of utility for this information on regional scales with account for uncertainty. Formulated are probabilistic measures of the effectiveness of monthly climate-model ensemble simulations of an indicator variable for discriminating high versus low regional observations of a target variable. The formulation uses the significance probability of the Kolmogorov–Smirnov test for detecting differences in the sample probability distribution of the conditioned high versus low regional observations of the target variable. Estimators that account for climate-model ensemble simulations and spatial association between the indicator and target variables are formulated. Generalizations for the cases of vector indicator and target variables are discussed. The methodology is exemplified for the case of a single climate-model indicator variable, seasonal surface precipitation, and a single regional target variable, seasonal mean areal precipitation over a U.S. climate division. Information from 10-member ensemble simulations of the German ECHAM3 atmospheric climate model for January 1950–December 1998 is used in the example, and results are presented for all the climate divisions of the conterminous United States. Monte Carlo simulation is used to establish the significance of the estimator values. The results show that the ensemble of climate-model seasonal precipitation simulations, when averaged over several model nodes, is skillful in discriminating the high from the low terciles of observed climate division seasonal precipitation in several regions of the United States and for all of the seasons. Over large regions in the southern, western, and northern United States in winter and spring, GCM simulations are likely useful for seasonal water resources studies on scales comparable to those of the climate divisions. The results for a coherent region east of the Rockies in summer and several regions of the northeastern United States and the southwest in autumn also exhibit significant potential benefits of using climate-model simulations for seasonal water resources studies.

Full access
Konstantine P. Georgakakos

Presented are necessary requirements of a modern-day flash-flood warning system that is capable of site-specific forecasts and that is suitable for national implementation. The requirements are identified based on the hydrometeorological character of the flash-flood phenomenon and on the real-time nature of the forecast procedure. Contemporary theories of heavy-rainfall and runoff generation and development are reviewed.

Full access
Dimitris Tsintikidis and Konstantine P. Georgakakos

Abstract

The focus of this paper is the elucidation of the physical origins of the observed extreme rainfall variability over tropical oceans. A simple statistical–dynamical model, suitable for use in repetitive Monte Carlo experiments, is formulated as a diagnostic tool for this purpose. The model is based on three partial differential equations that describe airmass, water substance, and vertical momentum conservation in a column of air extending from the ocean surface to the top of the storm clouds. Tropospheric conditions are specified for the model state variables (such as updraft–downdraft velocity, precipitation water and cloud content, or saturation vapor deficit) in accordance with past observations in oceanic convection, to allow for vertical integration of the model equations and the formulation of a computationally efficient diagnostic tool. Large-scale forcing is represented by stochastic processes with temporal structure and parameters estimated from observed large-scale data. This model formulation allows for sensitivity studies of surface rainfall temporal variability as it is affected by microphysical processes and variability in large-scale forcing. Dependence of the results on model-simplifying assumptions is quantified. Data from the Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment are used to validate the formulation statistically and to produce forcing parameters for the sensitivity studies. On the basis of Monte Carlo simulations that resulted in the generation of 10-min rainfall rates averaged over 4 km × 4 km, it is found that (a) the probability distribution function of model-generated rainfall resembles that of observed rainfall obtained by rain gauges and radar; (b) the power spectra of the model-generated rain time series, while reproducing the power-law character of the observed spectra for high rain rates, have generally steeper slopes than those of the radar-observed ones; (c) the character and magnitude of the model-generated rainfall variability are substantially influenced by the model microphysical parameterization and, to a lesser extent, by the shape of the vertical profiles of the state variables; and (d) while the probability of local rain is substantially influenced by both thermal buoyancy and water vapor availability, the exceedance probability of high rain rates (>10 mm h−1) is much more sensitive to changes in the former than in the latter large-scale forcing. The quantitative results of this work may be used to establish links between deterministic models of the mesoscale and synoptic scale with statistical descriptions of the temporal variability of local tropical oceanic rainfall. In addition, they may be used to quantify the influence of measurement error in large-scale forcing and cloud-scale observations on the accuracy of local rainfall variability inferences, important for hydrologic studies.

Full access
Jianzhong Wang and Konstantine P. Georgakakos

Abstract

A total of 62 winter-storm events in the period 1964–99 over the Folsom Lake watershed located at the windward slope of the Sierra Nevada were simulated with a 9-km resolution using the fifth-generation Pennsylvania State University–NCAR Mesoscale Model (MM5). Mean areal precipitation (MAP) over the entire watershed and each of four subbasins was estimated based on gridded simulated precipitation. The simulated MAP was verified with MAP estimated (a) by the California Nevada River Forecast Center (CNRFC) for the four subbasins based on eight operational precipitation stations, and (b) for the period from 1980 to 1986, on the basis of a denser precipitation observing network deployed by the Sierra Cooperative Pilot Project (SCPP). A number of sensitivity runs were performed to understand the dependence of model precipitation on boundary and initial fields, cold versus warm start, and microphysical parameterization. The principal findings of the validation analysis are that (a) MM5 achieves a good percentage bias score of 103% in simulating Folsom basin MAP when compared to MAP derived from dense precipitation gauge networks; (b) spatial grid resolution higher than 9 km is necessary to reproduce the spatial MAP pattern among subbasins of the Folsom basin; and (c) the model performs better for heavy than for light and moderate precipitation. The analysis also showed significant simulation dependence on the spatial resolution of the boundary and initial fields and on the microphysical scheme used.

Full access
Alexandre K. Guetter and Konstantine P. Georgakakos

Abstract

The association between the El Niño/La Niña and seasonal streamflow for the Iowa River is investigated. The seasonal Southern Oscillation index (SOI) was ranked and the extreme quartiles for each season were selected to condition the composite analysis of streamflow. The either concurrent or lagged association between anomalous SOI index and streamflow was obtained with a composite analysis that windowed a 3-yr period. The existence of statistically significant streamflow responses to El Niño and La Niña has been demonstrated for lags ranging from zero to five seasons. The long lag of streamflow-SOI association is attributed to 1) the time to establish global and regional circulation conducive to excess or deficit rainfall in the Midwest and 2) the inertia of anomalous high (low) soil water. Streamflow responses to El Niño and La Niña are out of phase. Above normal streamflow is associated with El Niño, whereas dry conditions are associated with La Niña. Sensitivity analysis of the streamflow-SOI association with respect to the magnitude of SOI seasonal anomalies suggests that winter SOI < −0.73 yields above normal streamflow from fall (three-season lag) to spring (five-season lag), with 70% consistency. Below-normal streamflow during fall is associated with SOI > 0.63 in preceding spring and summer, with 70% and 15% consistency, respectively. Streamflow predictive models conditioned on SOI anomalies were developed for lead times up to five seasons.

Full access
Nicholas E. Graham and Konstantine P. Georgakakos

Abstract

Numerical simulation techniques and idealized reservoir management models are used to assess the utility of climate information for the effective management of a single multiobjective reservoir. Reservoir management considers meeting release and reservoir volume targets and minimizing wasteful spillage. The influence of reservoir size and inflow variability parameters on the management benefits is examined. The effects of climate and demand (release target) change on the management policies and performance are also quantified for various change scenarios. Inflow forecasts emulate ensembles of dynamical forecasts for a hypothetical climate system with somewhat predictable low-frequency variability. The analysis considers the impacts of forecast skill. The mathematical problem is cast in a dimensionless time and volume framework to allow generalization. The present work complements existing research results for specific applications and expands earlier analytical results for simpler management situations in an effort to draw general conclusions for the present-day reservoir management problem under uncertainty. The findings support the following conclusions: (i) reliable inflow forecasts are beneficial for reservoir management under most situations if adaptive management is employed; (ii) tolerance to forecasts of lower reliability tends to be higher for larger reservoirs; (iii) reliable inflow forecasts are most useful for a midrange of reservoir capacities; (iv) demand changes are more detrimental to reservoir management performance than inflow change effects of similar magnitude; (v) adaptive management is effective for mitigating climatic change effects and may even help to mitigate demand change effects.

Full access
Alexandre K. Guetter and Konstantine P. Georgakakos

A record of 50 years of daily outflows through the boundaries of the continental United States has been assembled based on observations recorded by U.S. Geological Survey streamflow stations. Only stations with continuous records from 1939 through 1988 were included. These stations (197 total) are near the outlets of rivers located at the vicinity of the Canadian, Mexican, Atlantic (including the Gulf of Mexico), and Pacific borders of the continental United States. The drainage area of the selected stations covers 77% of the conterminous United States, whereas the existing network of gauging stations covers 83% of the conterminous U.S. area. Station daily data were aggregated over the entire boundary of the United States and were integrated in monthly and annual totals. The 50-year average annual streamflow divergence normalized by the aggregated drainage area is 210.2 mm yr−1 reaching a peak in April with 27.3 mm month−1 and a minimum in September with 8.7 mm month−1. The Mississippi–Missouri Basin comprises 50% of the gauged area and dominates the absolute value of the outflow discharge. Spectral analysis of the monthly outflow anomalies shows an 11-year dominant cycle. The 1939–1988 period contains four notable droughts. Two droughts are partially registered in the limits of the records characterized by the negative anomalies extending from 1939 to 1941 and by the 1987–1988 anomalies for the late 1980s drought. The middle 1950s and early 1960s droughts are fully included in the dataset. Periods of high outflows were registered in the middle 1940s, early 1970s, and early 1980s. Analysis of the spatial coherence of the annual anomalies shows large-scale features, whereas analysis of the monthly anomalies yields the frequency and persistence patterns of floods and droughts. An estimate of the climatological land-surface water budget for the continental United States was done based on recorded precipitation, panevaporation, and outflow. Eigenvector analysis of the monthly outflow residuals per 3° range has been performed to identify the major modes of the spatial correlation structure. The first eight modes explain 66% of the variance of the system and identify the following regions: Atlantic seaboard, Mississippi–Missouri and Ohio River basins, Northeast, Pacific Northwest, Pacific seaboard, Texas Gulf region, North-central, and the Colorado River and Great Basin. Annual and monthly specific outflow aggregates were used to describe the temporal characteristics of the coherent regions. Both time-domain and spectral analyses of the regional outflow anomalies identify the dominant modes of temporal variability.

Full access
Konstantine P. Georgakakos and Nicholas E. Graham

Abstract

This paper examines the conditions for which beneficial use of forecast uncertainty may be made for improved reservoir release decisions. It highlights the parametric dependencies of the effects of uncertainty in seasonal inflow volumes on the optimal release and objective function of a single reservoir operated to meet a single volume target at the end of the season under volume and release constraints. The duration of the “season” may be one or several months long. The analysis invokes the application of Kuhn–Tucker theory, and it shows that the presence of uncertainty introduces complex dependence of the optimal release and objective function on the reservoir parameters and uncertain inflow forcing. The seasonal inflow volume uncertainty is represented by a bounded symmetric beta distribution with a given mean, which is considered to be unbiased, and a half-range QR. The authors find that the use of predicted inflow uncertainty is particularly beneficial during operation with a volume target that is either near reservoir capacity or near zero reservoir volume, with the optimal release being directly dependent on QR in these situations. This positive finding is moderated by the additional finding that errors in the estimation of predicted QR can result in significant operation losses (larger deviations from the target volume) that are due to suboptimal release decisions. Furthermore, the presence of binding release constraints leads to loss of optimal release and objective function benefits due to the seasonal inflow uncertainty predictions, suggesting less rigid release policies for improved operations under uncertain forecasts. It is also shown that the reservoir capacity values for which optimal reservoir operations are most sensitive to seasonal inflow uncertainty predictions are found to be at most 5 times the uncertainty range of the predicted seasonal inflow volume and to be at least as large as the uncertainty range of predicted inflow volumes. Suggestions for continued research in this area are offered.

Full access
Konstantine P. Georgakakos and Michael D. Hudlow

Quantitative hydrologic forecasting usually requires knowledge of the spatial and temporal distribution of precipitation. First, it is important to accurately measure the precipitation falling over a particular watershed of interest. Second, especially for small watersheds and/or for longer forecast lead times, forecasts of precipitation are critical to the achievement of the greatest possible hydrologic forecast accuracy and longest possible lead time. This paper describes the current hydrologic forecasting program of the U.S. National Weather Service (NWS) and highlights the relevance of Quantitative Precipitation Forecasting (QPF) products to real-time hydrologic forecasting. Specific requirements for QPF products in support of hydrologic forecasting applications are defined and current operational QPF procedures are reviewed to determine to what extent they meet these requirements. It is concluded that no known QPF procedures capable of fulfilling all desired requirements are currently available operationally, although much valuable QPF information is available to meet parts of these requirements. Some recent advances in mesoscale QPF research are examined and these techniques are treated in two categories: those uncoupled dynamically from and those dynamically coupled to hydrologic forecasting procedures. Finally, a summary of possible future directions toward achieving improved use of QPF information in hydrologic forecasting applications is presented.

Full access