G. Padmanabhan

Analysis and modeling of Palmer's drought index series

Abstract The objective of the present study is to investigate the stochastic nature of yearly and monthly Palmers drought index (PDI) series and to characterize them via valid stochastic models which may be used to forecast and to simulate the PDI series. The monthly and annual PDI series for Iowa (1930–1962) and Kansas (1887–1962) are analyzed in the present study. The period of data includes the severe droughts experienced in U.S.A. in the 1930s. Valid autoregressive models are fitted to these time series and the final models are selected on the basis of class selection rules. Statistical characteristics of the observed data and those generated by the selected models are compared. The selected models are also tested for their forecasting ability. Results of these analyses and the modeling effort demonstrate that PDI series can be forecast with reasonable accuracy, one to several months ahead. The yearly forecasts are much less accurate. The models can also be used to generate synthetic data which preserve the important statistical characteristics such as the long-term oscillations of the original data.

APEX FEEDLOT WATER QUALITY SIMULATION

A manure erosion equation was developed and added to the APEX model for use in estimating nutrient losses from feedlots and manure application fields. The modified APEX was validated with data from feedlots near Bushland, Texas, and Carrington, North Dakota. The model was used to investigate feedlot management options on a hypothetical feedlot with realistic data. Vegetative filter strip (VFS) characteristics including 10%, 25%, 50%, and 100% FLRs (flow length ratios, i.e., filter flow length/feedlot flow length) and slopes of 1%, 5%, and 10% were considered. Combinations of these VFS characteristics were compared on two soil types: clay loam and fine sandy loam. Management options included three stocking rates (10, 15, and 20 m2 head-1) and two clean-out intervals (90 and 180 d). Additionally, two climatic conditions (precipitation of 440 and 825 mm year-1) were simulated. Results from 50-year simulations indicate that a VFS downslope of the feedlot can greatly reduce nutrient loads. All three VFS characteristics (FLR, slope, and soil) were important in controlling organic N and P losses. The best organic N and P control was obtained from a VFS with maximum FLR (100%), minimum slope (1%), and a sandy loam soil. Results were similar for soluble N and P control except that VFS slope had little effect. The simulated management options (clean-out interval and stocking rate) were also effective in controlling nutrient losses. The climatic variable (annual precipitation) gave higher nutrient losses from the feedlot and the VFS with 825 mm than with 440 mm. Nutrient control efficiencies, CEs, 100.* (1.0 - nutrient loss from VFS / nutrient loss from feedlot) were calculated for all scenarios considered. The VFSs on sandy loam soil with FLRs equal to or greater than 50% gave the highest CEs for both soluble and organic nutrients. Other factors including VFS slope, clean-out interval, and stocking rate had marginal impacts on CE. For soluble nutrients, CE is inversely related to annual precipitation. Thus, it is important to locate feedlots in areas with low precipitation and to provide a well designed VFS. The APEX model with the new manure erosion equation provides a tool for designing VFSs for controlling nutrient losses from feedlots.

Assessing the Influence of Wetlands on Flooding

The aftermath of the 1997 Red River of the North flood led to questions regarding the value of wetlands for flood control. The HEC-1 flood hydrology model was used to investigate the integrated effect of wetlands on flooding over a watershed scale. The Maple River watershed with a drainage area of 4150 km 2 in the Red River Valley was chosen for the study. The HEC-1 model was chosen because of its capability to: 1) simulate major hydrologic processes during a flood event, 2) subdivide the watershed to reflect spatial variations in parameters and locations of wetlands, 3) simulate wetland storage (either as reservoirs or diversions), 4) simulate watersheds with drainage areas exceeding 2500 km 2 , 5) use a variety of overland and streamflow routing methods, and 6) interface with GISs and DEMs for preparation of input data and post-processing of the results. Types and surface areas of wetlands were identified in each sub-watershed and storage volumes were estimated for the drained wetlands in each sub-watershed. The estimated storage provided by restoring these drained wetlands was modeled using diversions in the HEC-1 model. As flow arrives at the outlet of each sub-watershed, a percentage is diverted from the system, representing the runoff that might be held within wetlands. This method has the advantage of not assuming that all runoff will be intercepted by the restored wetlands. The rate of diversion can be varied so that all storage is sure to be utilized. The HEC-1 model was run within WMS (Watershed Management System) software, and output hydrographs were generated by the WMS. The results indicated that restoring all drained wetlands (representing about 0.24% of surface area in the Maple River watershed) would lower the flood stage for the 100 year flood event by less than 0.4%) even when the wetlands are modeled as empty before the flood event begins. Increasing the percent of restored wetlands to 1% by watershed area (4 times the present area) reduced the 100-year flood stage by only 0.9%. Therefore it does not appear that the volume of storage gained by restoring the drained wetlands in the Maple River watershed would significantly affect a low-frequency event such as the flood of 1997.

Toxicological evaluation of the leachate from a closed urban landfill

Landfilling is one of the commonly used methods of disposal of solid wastes in small and medium size municipalities. Leachate produced in the landfills can contaminate underground sources of water. Though precautionary measures to reduce the risk of leachate migration into underground sources of water are included in the selection of sites for landfilling and in their design and construction, the risk of contamination is not completely eliminated. Though chemical and in vitro toxicity tests on leachates have been conducted, not many studies are related to long-term toxicity of leachate in animal models. Therefore, toxicological evaluation of leachates is needed in order to predict possible undesirable effects of consumption of leachate-contaminated water on human populations. In this study, leachate obtained from a close landfill, which served a city population of about 60,000, was evaluated for its toxicity to mice. Animals were given leachate as drinking water for 65 days, and general toxicity to selected end points was observed.

Intervention analysis of total suspended particulate data from Chicago

Abstract The monthly total suspended particulate data measured at Chicago during January 1964 to December 1977 are analyzed and the results are reported. The ban on using high sulfur coal in Chicago which came into effect in January 1970 coincided with a dramatic change in the seasonality of the particulate time series and a reduction in particulate levels. Traditional methods of time series analysis based on the assumptions of constant mean and variance of time series are not applicable for these nonstationary time series with jumps. A recently developed stochastic modeling technique called intervention analysis is used in the present study. A dynamic model and a stochastic model are used to characterize the time series. The technique is briefly introduced and the results of analysis of the suspended particle data are discussed. The statistically significant reduction in the suspended particles in Chicago after January 1970 has occurred over both the warm and cold periods of the year. The dangers of using a model fitted to data before the intervention in the post-intervention period are also brought out.

Water Quality Modelling of a Proposed Reservoir.

This chapter introduces modelling techniques that have been found to be successful in evaluating future water quality conditions in proposed reservoirs. An ex…

Revisiting steady state water flow in a saturated, inclined soil slab

Mathematical models of flow in a saturated, inclined soil slab with an impermeable lower boundary have contributed substantially to understanding of important processes in the environment. One such process is movement of chemicals induced by interflow in sloping soil. Certain aspects of previous mathematical treatment regarding the inclined soil slab problem have, however, escaped scrutiny. The original 1965 treatment includes a peculiar quadratic boundary condition term that applies over a vanishingly small fraction of the soil surface. The quadratic term is shown to be unjustified and unnecessary in obtaining practical solutions. Mathematical solutions of two more general inclined slab geometries, three-dimensional and parallelogram slabs, are presented and solved. The solutions are illustrated in terms of figures showing selected flow paths and equipotentials.

Heuristic Knowledge-based Tool for Rainfall Synthesis, Runoff Estimation and Hydrograph Generation

Knowledge-based engineering has emerged as a potential technique for incorporating human expertise and some degree of intelligent judgment into decision-supporting procedures. A knowledge-based expert system (KBES) methodology to estimate the volume and hydrograph of direct surface runoff from rain events using the Soil Conservation Service Runoff Curve Number method was developed for hydrologic modeling and decision support systems. The KBES approach was designed to determine runoff curve numbers with limited available information on watershed characteristics, to generate rainfall intensity of the geographic location for desired durations and return periods, and to use the curve number and rainfall intensity estimates to calculate runoff volume, peak runoff and time-to-peak for design purposes. Rainfall intensity estimation is based on four regional partial-duration series parameters corresponding to a given geographic location. Currently, a regional parameter database for 12 midwestern states (Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin) is compiled into the KBES for the rainfall synthesis.