H51J-01 INVITED
Monitoring and Modelling the Influence of Overland Flow on Hydrology and Hydrochemistry at Different Scales in Montane Watershed
Overland flow is often a significant hydrological process in montane watersheds; however, its contribution to quantity and quality of stream flow during storm responses depends upon factors such as soil cover, topography and precipitation characteristics. Recent work in 18 contrasting mesoscale catchments in the Scottish highlands compared input-output relationships for natural conservative tracers and hypothesised that greater tracer damping largely reflects the diminishing importance of overland flow and other near-surface hydrological flow paths. This hypothesis is supported by very good predictive relationships between the percentage cover of soils likely to generate overland flow (histosols, regosols, sealed surfaces etc.) and tracer-derived descriptors of catchment-scale hydrological function such as mean residence time and annual percentage groundwater contribution to stream flow. Put simply, the greater the coverage of overland flow-dominated soils, the shorter the mean residence times and the lower the groundwater contributions. In this contribution, the hypothesis is further tested by analysis of tracer data and digital soils maps) from the UK Hydrology of Soil Types (HOST) system) of 20 different validation catchments varying between 1km2 and 240km2. Moreover, it is hypothesised that stronger predictive relationships might be derived by considering the distribution of riparian soils that are highly connected to the river channel network. Preliminary results indicate that, in general, the relationships hold; though riparian soil cover is no better a predictor of tracer damping than total catchment soil cover. This may, however, reflect inadequacies in the readily available soil maps or the need for a more sensitive, topographically-based definition of riparian areas. The results of these field campaigns are now being integrated in tracer-aided, semi-distributed hydrological models. The modelling work confirms that fundamentally different mixing relationships need to be considered in catchments where overland flow is a dominant mechanism of storm runoff generation.
H51J-02
Overland flow caused by groundwater springs in the Catskill Mountains, New York, USA
Groundwater springs have been shown to be important for baseflow maintenance, ecological diversity, and biogeochemical transport in the Catskill Mountains of New York State. A study was undertaken to evaluate the importance of groundwater springs to stream flow response, spatial distribution of saturated areas, and chemical transport on a 2 km2 watershed. Discharge and water chemistry were monitored at five upland springs and the watershed outlet during three storm events and weekly for one year. The spatiotemporal connectivity of groundwater springs to the stream was evaluated using the isotopes, conservative geochemical tracers, groundwater heights, temperature probes, and soil moisture measurements. The data indicates that spring discharge is capable of maintaining isolated areas of saturation and overland flow that provide hydrologic connections between upland hillslopes and the stream. Thus, upslope springs can be important sources of solutes and nutrients, especially during baseflow periods. The response of the individual springs was correlated to the upslope contributing area during large rain events. However, terrain indices were not capable of predicting relative spring discharge during moderate to dry conditions. The results suggest that two geomorphologic wetness states exist for this Catskill watershed: 1. during wet antecedent conditions surface topography is a first- order control on water table heights and overland flow, 2. during dry to average antecedent wetness conditions geomorphologic heterogeneities (e.g. bedrock fractures and confining layers) control locations and extent of saturated areas and overland flow. The two-state wetness system hypothesis has important ramifications for developing watershed-scale parameters (i.e. drainage density) and spatial modeling in heterogeneous landscapes.
H51J-03
Saturation Overland Flow in Forests? A Model Evaluation
Overland flow generated from saturated areas contributes substantially to storm runoff in many humid watersheds. Several rainfall-runoff models have incorporated saturation excess overland flow (SOF) as an important runoff generation mechanism. This includes our Sensitive Area Mapping Model (SAMM), which has been applied to the entire province of British Columbia (BC, Canada) to predict the effects of changes in forest management practice on hydrology. SAMM, and in particular its SOF module, is evaluated with field studies in forested watersheds with different precipitation regimes throughout BC. To detect saturated areas, five sampling methods were tested and compared: (1) Saturation occurrence samplers were applied; (2) temperature loggers installed to detect delays in soil temperature changes when saturation occurs; (3) SOF controlling morphology characteristics were delineated; (4) wetness indicator plants registered; and (5) soil profiles were analyzed. Our research demonstrates that all five methods can be applied to identify SOF generating areas. The saturation occurrence samplers and temperature loggers provide a good temporal resolution detecting SOF, whereas the other methods are based on long-term averaged characteristics. The assessment of morphology, indicator plants and soil profiles (MIS) can be applied to entire watersheds, whereas the other two methods are limited to smaller areas. Consequently, saturation occurrence samplers and temperature loggers were installed in a research forest to compare the temporal variability of SOF with the predictions of SAMM. Furthermore, the MIS methods were applied to evaluate the spatial variability of SOF in four catchments of different climate (Pacific coast/wet, mainland/moderate, Okanagan/dry, Rocky Mountains/alpine). In each basin more than a hundred forested 25m grid cells were sampled and analyzed. The combination of morphology, indicator plants, and soil profile data allowed estimating several indices which all express each grid cell's disposition to generate SOF. These indices are compared to our model's saturation prediction with rank correlation tests and Kappa statistics. Our research provides valuable information about the relevant processes that control SOF in forested ecosystems with different precipitation regimes. Furthermore, our study presents methods to evaluate SOF modules of hydrologic models. This diagnostic tool helps us to select models which describe the internal characteristics of a watershed appropriately.
H51J-04
Hortonian Overland Flow in Humid Forested Environments: Occurrence, Measurement, Characterization, and Importance
Minimization, dispersal, and filtering of Hortonian overland flow is a main focus of forestry Best Management Practices. One of the several water quality functions of a forested riparian buffer is to filter surface runoff from upslope. However, few studies have quantified either the temporal frequency, spatial extent, or quantity of Horton overland flow from mature forests, clearcuts, or young plantations. As part of a multi-phase study of overland flow from silvicultural operations, we installed overland flow collector cups every 30 meters along the perimeter of streamside management zones of two clearcuts in the Georgia Piedmont, and we also installed overland flow collector cups along the imaginary SMZ boundary of a forested reference watershed. The cups were checked and emptied after every storm greater than 13 mm. Differentiating between Horton overland flow and saturated surface flow was difficult. Some responsive cups were clearly located in variable source areas as indicated by vegetation, soils, or the tendency of the cups to float, and other responsive cups were located in places where saturated surface flow would not be expected. For many responsive cups, however, differentiation between overland flow mechanisms was ambiguous. Since both saturated surface flow and Horton overland flow will concentrate due to topographic convergence, location cannot differentiate the flow mechanisms. Coincident measures of shallow groundwater are necessary to define surface flow generation mechanism. On all sites, the number of cups responding increased with the depth of the storm. The spatial distribution of responsive cups was only partly explained by topography. As expected, overland flow was common on the clearcut areas, and became less common over time as vegetation re-established. Overland flow was less common on the forested reference site, but still frequent and sometimes spatially extensive.
H51J-05
Factors Influencing Surface Runoff and Hydrologic Connectivity on an Agricultural Hillslope in Central Pennsylvania
Improved understanding of surface hydrologic processes is central to the targeted application of agricultural management practices for water quality protection. Factors influencing surface runoff production and hydrologic connectivity were explored at three landscape positions on a single hillslope located within a 27.4 acre agricultural watershed in central Pennsylvania. A total of 119 events (rainfall, rain-on-snow, and snowmelt) were monitored for total runoff volume (mL), water-table depth (0 cm - 45 cm), and soil-moisture content from 2002 to 2004 using 1- by 2-m plots at seepage slope, transportational midslope, and colluvial footslope positions. The seepage slope and transportational midslope plots were located in well-drained residual soils, whereas the colluvial footslope plots were located nearest the stream channel in somewhat poorly drained soils with clay fragipans. Poorly-drained soils in the colluvial footslope position generated more frequent (> 2 times) and significantly higher volumes of runoff (> 19 times) than the well-drained soils in the transportational midslope or seepage slope positions. Runoff generation on the colluvial footslope was predominately associated with elevated water-table depths and occurred most frequently during the late fall and winter months. In contrast, runoff generation on the well-drained seepage and transportational midslopes occurred most frequently during the summer months and was predominately associated with lower water-table depths (i.e. infiltration-excess surface runoff). Hydrologic connection, defined herein as runoff occurring simultaneously on the upslope and downslope plots, was observed for 39% of the events and occurred most frequently during the spring and early summer periods. The overall runoff trends observed in this study illustrate the importance of considering seasonal and landscape variables in agricultural watershed management.
H51J-06 INVITED
Overland Flow Generation Under Contrasting Land use in Amazonia (Rondônia, Brazil)
Human transformation of the Earth's land surface has large and important consequences for the functioning of hydrological processes in watersheds. Changes in land cover and land use can modify the relative contribution of overland flow, subsurface flow and groundwater flow to stream discharge. The conversion of tropical primary forest to pasture is known to reduce soil hydrologic conductivity (Ksat) significantly. To study the influence of reduced Ksat on overland flow generation and to get insight into the interaction of Ksat, the development of perched water tables and the occurrence of overland flow, we compared the hydrology of two small catchments under primary open tropical forest and pasture in the southwestern Brazilian Amazon basin. Of 176 rainfall events that occurred from August 2004 to July 2005, 70 events generated stream flow in the pasture, but only 37 events generated stream flow in the forest catchment. Overland flow was 18% of incident rainfall in the pasture but only 1% in the forest. On event basis, the overland flow response ranged between 0 and 54% of incident rainfall in the pasture and 0 and 14% of incident rainfall in the forest. Piezometer with automatic water level loggers were installed in 12.5, 20 and 50 cm soil depth at ten sites per catchment. Perched water tables were observed in 12.5 and 20 cm soil depth in both catchments, but more frequently in the pasture. Piezometer hydrographs depended on rainfall characteristics and on their position relative to the stream channel and indicate slow subsurface flow. We conclude from our study that reduced Ksat due to land use change from primary forest to pasture results in 1) a significant increase of overland flow and 2) a more frequent shift between aerobic and anaerobic conditions from the soil surface to a depth of 20 cm. We expect both phenomena to influence catchment biogeochemistry, as illustrated by a closer look at nitrate.
H51J-07
Significance of Overland Flow in Sustaining Water Resources of Arid and Semi-Arid Rivers - Water Quantity and Quality Implications
Overland flow is known to be a dominant runoff generation mechanism in arid and semiarid river systems. Despite its prevalence, little is known about the impact of overland flow on the quantity and quality of water in arid and semi-arid rivers and riparian systems. Several studies along the San Pedro and Rio Grande Rivers in the Southwest United States have documented the importance of ephemeral overland flows to the quantity and quality of river water in the stream and near stream zones. First, studies on both rivers have documented the importance of flood flows in providing a significant source of water to near stream aquifers. On the San Pedro River studies have shown that ~50% of baseflow water originates from summer Monsoon floods with a stronger influence on losing versus gaining river reaches. In the Rio Grande, stable isotope data indicate that nearly 100% of the increase in discharge during a Monsoon flood event can be attributed to ephemeral overland flow with approximately 40% of this flood pulse in the Rio Grande lost to the shallow alluvial aquifer. Second, nutrient studies on both rivers demonstrate that reconnecting the river with its uplands during flood events causes a dramatic increase in nutrient concentrations and fundamentally alters near and in-stream biogeochemical conditions and processes by providing a large pulse of allochthonous nutrients and organic matter. Despite the large nutrient influx with flood events the sustained impact on nutrient composition is limited; with upwelling zones, possibly reworking particulate organic matter, having more influence than water source on in-stream nutrient concentrations. Furthermore the influence of flood events attenuates significantly over a period of months with the influence of flood events diminishing from 40% of river flow to 20% in a period of just 6 months. http://hwr.arizona.edu/tmeixner
H51J-08
Overland Flow in Forests: Not Seeing the Soil for the Trees
The regional geography approach to forest hydrology and a historical focus of hydrologic process research on temperate forests has resulted in an uncritical identification unsupported by global data of forests in general with temperate forests regarding runoff generation. Overland flow, in particular, a flowpath considered a rare occurrence in temperate forests, has become all but incompatible with the term "forest", a few admissible (e.g. "localised") exception notwithstanding. While this conclusion - "forest, therefore no overland flow" - may be dismissed as a case of induction gone awry, it has had, in the sense of a self-fulfilling prophecy, an insidious effect on field methods: why bother to detect or monitor overland flow in an ecosystem that almost by definition precludes its existence? This presentation therefore strives 1) to separate hydrologically relevant soilscape properties from methodological inadequacies in an attempt to re-assess the global relevance of overland flow in forests, and 2) to explore to which extent these methodological inadequacies derive from an outdated understanding of overland flow-generating mechanisms in forests. Absence of evidence is not evidence of absence.