Sediment Modeling

doten

Preliminary application of the DHSVM sediment module

Introduction

Water and land resource management decisions have a direct effect on the environmental and economic sustainability of a watershed system. One of the major consequences of unchecked land use change is the impact on water supply and quality, downstream hydrological hazards (e.g. floods) and biodiversity conservation (Dunne, 1997; Tong et al., 2002; Schroter et al., 2005; Bates et al., 2008). Alteration of the landscape and other human-caused disturbances has been shown to be important factors affecting mass transport (loading) of principal plant nutrients (nitrogen and phosphorus) and sediment to   lakes (Loeb, 1988). 
Rapidly developing countries, particularly those evolving from rural to urban, frequently lack the resources to adequately monitor and forecast the impact of land use changes on biologic, hydrologic and social resources. This leads to a growing need to develop a comprehensive set of tools, capable of accurately predicting the impact of land use changes on small to large spatial scales  (Defries et al., 2004). Accurate estimation of mass and chemical export from a watershed is needed to predict the biogeochemical response of a system to changes in both land use and climate (Hope et al., 1994; Tranvik et al., 2002).
The Santa Maria da Vitória watershed provides 30% of the potable water to the capital city of Vitória, Espírito Santo state, in Brazil. It covers an area of roughly 2000 km2, with a river length of 122km. Mountainous landscapes and intense whitewater rapids characterize the upper region of the Santa Maria da Vitória basins. The landscape of the Santa Maria da Vitória watershed is composed of roughly 40% natural forested land cover,   mostly concentrated in the upper regions of the basins. The SMV passes through five municipalities of varying size: Santa Maria de Jetiba, Santa Leopoldina, Cariacica, Serra, and Vitória. The SMV river mouth empties into the Bay of Vitória, an important commercial port for the region, and forms the island of Vitória. The river basin is   influenced by rural development, pasture land, and agricultural export plantations such as banana, coffee, eucalyptus, and sugar cane.


Despite the importance of the river basin to the state, land use change has resulted in sediment load increase to the Bay of Vitória and upstream dams predominantly during the rainy season (Fig. 4.1). Sediment concentrations reach such high levels that reservoirs become inoperable, causing a decrease in water supply. There is little to no treatment of wastewater in the upper and central regions of the basin. Since the sewage treatment in the region uses septic tanks, septic tank outfall has provided the river with enriched organic nitrogen values. In the lowlands, agricultural activity has caused an increased in fertilizer activity within the SMV river.
The increase in sedimentation and chemical pollutants to the river has caused tension in the local populace, chiefly because the citizens do not understand the correlation between heavy rainfall and water scarcity. This has motivated the government of Espírito Santo to explore tools to quantify sediment fluxes and dynamics in the SMV watershed and to develop suitable responses for improving water quality and sanitation.


A major challenge however, is quantifying sediment dynamics – both magnitude and processes – in a region with very little data.   In this section we will examine the feasibility of the Distributed Hydrology Soil Vegetation Model’s (DHSVM)’s prototype sediment module (Doten et al., 2006) to describe sediment yields in the basin. 

 



Methods

    Doten et al. (2006) developed a sediment algorithm for DHSVM by using the existing slope stability models the Level I Stability Analysis model (LISA) developed by Hammond et al. (1992), and the European Hydrological System sediment (SHESED) model (Burton and Bathurst 1998, Wicks and Bathurst 1996). The algorithm redistributes sediments by computing the basin hydrology using overland flow, runoff and mass failure of sediment (Fig. 4.1).  DHSVM models the sediment loads over the entire basin for each time period. The variables used in sediment computations are: depth to saturation, saturation and infiltration excess runoff, precipitation, leaf drip and channel flow.

   The application of DHSVM to evaluating water flow conditions in the Santa Maria da Vitória River was described in Chapter 3. Here we examine the performance of the Doten et al (2006) sediment module. In this experiment, we only took into account the hillslope erosion and channel routing components, since this region is not particularly steep, thus reducing the need to compute sediment loads from landslides (mass wasting). The hillslope erosion component is not stochastic in nature, so it does not respond to each of the MASSITER failure scenarios predicted by the mass wasting component.  Hillslope erosion is also computed at the resolution of the hydrology model (in this case 10-m) and does not use the mass wasting model parameters. The sediment load due to hillslope erosion does not take into account additional sediment deposit from mass wasting events but is a product of overland flow (using the kinematic wave approximation) into the stream and channel segments.

   In order to run the sediment module for DHSVM in the SMV, a 10-m digital elevation model (DEM) and 10-m mask for the SMV was developed. Soil parameters and variables used were similar to those used in the land cover scenarios for Chapter 3. The landcover scheme was from IEMA2007.  For the meteorological data, we used the daily station data for precipitation, temperature, wind and humidity and disaggregated the forcings to 6-hourly timesteps.  This is the method outlined by Maurer et al, 2002 for the Variable Infiltration Capacity model (VIC). Since sediment yields can change drastically over a short period, we used 6-hour timesteps to capture any sudden events of erosion or mass wasting.

   The Distributed Hydrology-Soil-Vegetation Model (DHSVM, described elsewhere on this website) includes a sediment module. This module predicts the range and variability of catchment sediment yield due to mass wasting, forest roads surface erosion and overland flow in response to dynamic meteorological and hydrologic conditions. The DHSVM sediment module consists of four primary components: mass wasting, which is stochastic in nature; hillslope erosion; erosion from forest roads; and a channel-routing algorithm.  DHSVM provides a continuous temporal sequence, spatially distributed over a watershed, of the following variables used in the sediment computations: depth to saturation, saturation and infiltration excess runoff, precipitation, leaf drip and channel flow.


     DHSVM simulation for sediment yield at the SMV gauge C003 indicated that DHSVM produces downstream sediment loads that lags behind actual measured loads and also behind precipitation. (Fig. 1). One problem may be that spin-up or initialization time was insufficient to capture the basin sediment amount and dynamics since sediments in rivers are cumulative product. However, there was insufficient data to perform trend analyses or to draw conclusions with confidence.

Figure 1: Comparison between observed and simulated Total Suspended Solids (TSS) at SMV Gauge C003

 

    We also calculated the sediment yield using the power-law formulation for rating curves. The rating curve parameters were obtained for two different time-periods in order to reduce the slope of the linear regression line or the rating exponent, b. Initial analysis using the sediment rating curve method appears to produce a muchhigher load of sediments in the SMV than using DHSVM. The November 2013 sediment measures indicates that the SCR may show more promise in predicting sediment yield (see Fig. 1) since the sediment load predicted from the SCR follows precipitation (storm) vents more closely and produces variability similar to the measure loads during light rain periods; however, it should be noted that sample size of available measurements and streamflow is not sufficiently large enough to predict sediment loads with high confidence.

 


 

Future Work


    The Rio Mangarai is a tributary to the Rio Santa Maria da Vitoria, with considerable agriculture and high sediment loads, and "flashy" response to rainfall events.  Due to the small size of the Mangarai basin (about 1200 km2), the sediment load computation time should be reduced when using DHSVM. The Mangarai basin is also currently the site of a concerted effort to collect sediment, flow and other water quality measurements. Syvitski et al., 2000 points out that sediment rating coefficients are typically estimated with large errors even from nearby monitoring stations that have similar climatic and geologic features. This suggests that in order to fully capture sediment dynamics for a watershed, observations need to be collected on at least a daily level for individual rivers.


    This increase in field measurement data at a fine temporal and spatial scale would allow the development of SCRs with higher confidence. Our analysis above indicates that DHSVM and SCRs can and should be used in tandem to predict sediment yield in order to fully understand basin dynamics and address issues related to the spatial extent of erosion and soil destabilization since neither model nor SCR are “perfect”. The future work stemming from this chapter would also assess the effects of land use change (scenarios described in Chapter 3) on sediment loads in the SMV and Rio Mangarai; the results of which would be used to develop policies governing agriculture, riverine vegetation and evaluating the best areas for increased biodiversity.