Faculty of Engineering Kirikkale University Yahsihan Kirikkale Turkey 2 Pratt School of Engineering Duke University Durham NC USA

20.1 INTRODUCTION

The connections between climate, water resources, and hydrologic extremes as described by streamflow anomalies have been widely addressed in the literature (Schwarz 1977; Meyer et al. 1996; Kaczmarek et al. 1996; Hamlet and Lettenmaier 1999; and many others). However, the water resources impacts of climate variability go beyond changes in river water and indeed extend to all pathways of water in natural systems. Recently, increasing attention has been placed on the linkages between the so-called colors of water resources and sustainable development (Falkenmark etal. 1998). In this framework, rainfall is white water, streamflow (and groundwater) is blue water, and soil water content in the unsaturated zone is green water. Green water is the key water resource for food production (and ecosystem preservation), where irrigation is an artificial process of transferring blue water to green water. The underpinning rationale is that quantitative understanding of the dynamic balance between white, green, and blue water (i.e. the water cycle proper) is required to assess the long-term availability and resiliency of water resources at the spatial and temporal scales relevant for the human enterprise (Falkenmark 1997). Following this lead, our overarching objective is to assess the utility of hydrologic models to investigate the relationships among precipitation, vegetation, soil moisture dynamics and the intraseasonal variability of runoff production in complex landscapes. Specifically, we focus on elucidating the spatial and temporal variability of water and energy fluxes through interpretive analysis of model simulations of extreme hydrologic regimes.

Natural watersheds are characterized by high spatial heterogeneity that reflects interactions among the topographic, geomorphological, and biophysical characteristics of the landscape. The representation of spatial variability and associated nonlinear scale effects in hydrologic models remains a challenge both from the observational and from the modeling point of view (e.g. Binley et al. 1991; Seyfried and Wilcox 1995; Devonec and Barros 2002; among others). Our working hypothesis is that the utility of land hydrology models depends on their structural stability. By structural stability, we mean here the model's ability to simulate the onset and persistence of a hydrologic regime with consistent physics and without need for calibration (which is forcing dependent) or adjustment of model parameterizations. Structural stability implies predictive skill, and model portability or transferability (Barros 1995; Bindlish and Barros 2000; Devonec and Barros 2002). Further, we propose that structural stability can only be achieved when the spatial resolution of the model can resolve the fundamental spatial scales of the climate forcing

Climate and Hydrology in Mountain Areas. Edited by C. de Jong, D. Collins and R. Ranzi © 2005 John Wiley & Sons, Ltd

(i.e. rainfall) as well as landscape controls of hydrologie response (terrain, geology, soil, vegetation, etc).

For this work, a distributed Land Hydrology Model (LHM-3D) including soil-vegetation-atmosphere interactions, subsurface and overland flow, and streamflow (channel) routing is used for simulating warm season (spring and summer) hydrology during the 1988 drought and 1993 floods in the Monongahela River basin. The LHM-3D is a spatially distributed model that consists of three components: a one-dimensional (vertical column) Land Surface Hydrology Model (LSHM, Devonec and Barros 2002); a two-dimensional Surface Flow and River Routing Model with spatially and temporally varying routing parameters (SFRM; Yildiz 2001); and a two-dimensional Lateral Subsurface Flow Routing Model (LSFRM; Yildiz 2001). The LSHM simulates all processes involved in energy transfer between the atmosphere and the land surface (i.e. radiative fluxes, and ground, sensible and latent heat fluxes) as well as water transfers at the surface and within the unsaturated zone (snow processes, evapotranspiration, infiltration, and percolation) using spatially varying parameters and geomorphologic characteristics. Vertical and lateral subsurface flows are simulated sequentially by the LSHM

and the LSFRM, respectively. Groundwater divides are assumed to correspond to DEM-derived basin boundaries. In the Subsurface Flow Routing Model (SFRM), streams are allowed to interact with the land margins and the basin's groundwater system, and therefore river reaches may function as either gaining or losing streams. A detailed description of the model can be found in Yildiz (2001).

The Monongahela basin, a tributary of the Ohio River on the western flanks of the Appalachian Mountains, USA, is characterized by complex terrain and highly heterogeneous geology and vegetation cover and therefore provides an excellent setting to meet the objectives of our study (Figure 20.1). Sensitivity of physical processes to climate forcing (drought year of 1988 vs wet year of 1993) and the structural stability of the model as a function of spatial resolution (1 km vs 5 km) were investigated through analyses of the spatial and temporal variability of water fluxes across the basin, including precipitation, evapotranspiration, and runoff. The utility of the LHM-3D for climate impact studies on water resources was assessed by investigating the consistency between observed and simulated soil moisture (green water) dynamics across the basin, runoff

♦ Raingauge O Reservoir A Power station + Streamgauge

Figure 20.1 Digital Elevation Model (DEM) of the Monongahela River basin including raingauge and reservoir locations, stream network, and delineated subbasins with streamgauges at the outlets

♦ Raingauge O Reservoir A Power station + Streamgauge

Elevation (m)

Figure 20.1 Digital Elevation Model (DEM) of the Monongahela River basin including raingauge and reservoir locations, stream network, and delineated subbasins with streamgauges at the outlets production (blue water), and streamflow statistics, which has implications for reservoir and land management.

20.2 CASE STUDY: THE MONONGAHELA RIVER BASIN

0 0

Post a comment