WSPC07148 - Laserfiche WebLink

Home Browse Search

> , (i) the first order moment approach, in which the uncertainty is ex- pressed in terms of a variance associated with a variable of interes~. The variance is a complete descriptor of the uncertainty only in the case of guassian random variables. Nevertheless, the first-order second moment approach has been used with success in real-time forecasting (e.g. Georgaka- kos et al. 1988). In that study, optimal estimation of models states based on a physically based parameterization of uncertainty was shown to signifi- cantly improve the reliability of short-term forecasts. (ii) Monte-Carlo based approaches such as the Extended Streamflow Pre- diction methodology (Day 1985) which use a population of future meteorologi- cal records determined statistically based on records from previous years, together with a fixed set of initial conditions, for assessing long-term variability. Such a simulation framework has the advantage of generating a whole distribution (rather than just the first two moments). The long-term forecasts obtained in this approach are particularly valuable in the context of designing flood control structures for preparedness and resiliency. While these techniques have been used with lumped parameter watershed models, they have not been used extensively with distributed models in small watersheds prone to flash flooding. Distributed hydrologic modeling is becoming more feasible due to the ready availability of terrain and spatial data at high resolutions and distributed rainfall forecasts for systems such as NEXRAD (Beven 1991). Assessing the computational requirements associated with such detailed modeling and developing approaches for efficient implemen- tation of distributed modeling approaches remains a major challenge. Howev- er, in small watersheds with high relief, prone to flash flooding, such distributed modeling can significantly improve our understanding of flood dynamics. 5.3. Impact of Urban Water Conservation on Evapo-transpiration The state of the art in urban water conservation planning and management is summarized in an AWWA publication titled Evaluatin9 Urban Water Conserva- tion Programs: A Procedures Manual (Dziegelewski et al. 1993). We recently completed the first phase of doing continuous monitoring of residential water use in Boulder (Mayer 1995). Innovative, non-intrusive battery powered data loggers equipped with magnetic sensors were fitted on the water meters at 16 pre-selected homes in the Heatherwood neighborhood in Boulder and the flow rate was measured every 10 seconds during Summer 1994 and Winter 1995 (Mayer 1995). This yielded continuous flow trace data precise enough to isolate, quantify, and categorize individual water uses in each home. This database of over 15,000 measured events will be greatly expanded next summer with a study of 60 homes in Westminster, a northwest suburb of Denver. This expand- ed database will provide a unique capability to evaluate urban water supply systems. Also, for Heatherwood, we have established a complete GIS system for 220 houses which includes five years of measured monthly water use for each house and complete demographic information (Mayer 1995). Our studies to date indicate that nearly 80 % of summer water use is for lawn watering. Thus, this is potentially the most cost-effective control measure. What are the long-term implications of aggressive programs to reduce E-T from lawn watering in Boulder? An urban water budget will be developed as part of this initiative. Little work has been done to quantify the amount of water lost to ET in urban areas. In order to close the water budget for urban areas of the BCW, it is proposed to implement several methods to account for ET. Some simple methods such as the Penman and Priestley-Taylor equations adapted for vege- 11