Rohan Benjankar

Development of a spatially-distributed hydroecological model to simulate cottonwood seedling recruitment along rivers.

Dam operations have altered flood and flow patterns and prevented successful cottonwood seedling recruitment along many rivers. To guide reservoir flow releases to meet cottonwood recruitment needs, we developed a spatially-distributed, GIS-based model that analyzes the hydrophysical requirements for cottonwood recruitment. These requirements are indicated by five physical parameters: (1) annual peak flow timing relative to the interval of seed dispersal, (2) shear stress, which characterizes disturbance, (3) local stage recession after seedling recruitment, (4) recruitment elevation above base flow stage, and (5) duration of winter flooding, which may contribute to seedling mortality. The model categorizes the potential for cottonwood recruitment in four classes and attributes a suitability value at each individual spatial location. The model accuracy was estimated with an error matrix analysis by comparing simulated and field-observed recruitment success. The overall accuracies of this Spatially-Distributed Cottonwood Recruitment model were 47% for a braided reach and 68% for a meander reach along the Kootenai River in Idaho, USA. Model accuracies increased to 64% and 72%, respectively, when fewer favorability classes were considered. The model predicted areas of similarly favorable recruitment potential for 1997 and 2006, two recent years with successful cottonwood recruitment. This model should provide a useful tool to quantify impacts of human activities and climatic variability on cottonwood recruitment, and to prescribe instream flow regimes for the conservation and restoration of riparian woodlands.

Floodplain forest succession reveals fluvial processes: A hydrogeomorphic model for temperate riparian woodlands

River valley floodplains are physically-dynamic environments where fluvial processes determine habitat gradients for riparian vegetation. These zones support trees and shrubs whose life stages are adapted to specific habitat types and consequently forest composition and successional stage reflect the underlying hydrogeomorphic processes and history. In this study we investigated woodland vegetation composition, successional stage and habitat properties, and compared these with physically-based indicators of hydraulic processes. We thus sought to develop a hydrogeomorphic model to evaluate riparian woodland condition based on the spatial mosaic of successional phases of the floodplain forest. The study investigated free-flowing and dam-impacted reaches of the Kootenai and Flathead Rivers, in Idaho and Montana, USA and British Columbia, Canada. The analyses revealed strong correspondence between vegetation assessments and metrics of fluvial processes indicating morphodynamics (erosion and shear stress), inundation and depth to groundwater. The results indicated that common successional stages generally occupied similar hydraulic environments along the different river segments. Comparison of the spatial patterns between the free-flowing and regulated reaches revealed greater deviation from the natural condition for the braided channel segment than for the meandering segment. This demonstrates the utility of the hydrogeomorphic approach and suggests that riparian woodlands along braided channels could have lower resilience than those along meandering channels and might be more vulnerable to influences such as from river damming or climate change.

Comparison of Field-Observed and Simulated Map Output from a Dynamic Floodplain Vegetation Model Using Remote Sensing and GIS Techniques

Three different forms of the Kappa statistic were used to compare simulated and observed maps to assess the predictive capability of a dynamic vegetation model. Kappas were used to compare vegetation model performance considering both individual vegetation classes and merged vegetation phases. The overall Kappa increased slightly from 0.18 to 0.21 and 0.23, when different vegetation phases were resampled from the original 10 m cell size to 30 m and 50 m cell sizes, respectively. Similarly, overall Kappa improved from 0.18 to 0.55 and 0.71 when the neighborhood cells within a buffer distance of 10 m and 20 m, respectively, were considered. Different forms of the Kappa statistic (K, Kloc, Kqty) are valuable indicators to study the performance of vegetation models and identify parameters (e.g., quantity or location) for improvement in the model results.

Estimation of daily stream water temperatures with a Bayesian regression approach

&NA; Stream water temperature plays a significant role in aquatic ecosystems where it controls many important biological and physical processes. Reliable estimates of water temperature at the daily time step are critical in managing water resources. We developed a parsimonious piecewise Bayesian model for estimating daily stream water temperatures that account for temporal autocorrelation and both linear and nonlinear relationships with air temperature and discharge. The model was tested at 8 climatically different basins of the USA and at 34 sites within the mountainous Boise River Basin (Idaho, USA). The results show that the proposed model is robust with an average root mean square error of 1.25 °C and Nash‐Sutcliffe coefficient of 0.92 over a 2‐year period. Our approach can be used to predict historic daily stream water temperatures in any location using observed daily stream temperature and regional air temperature data.

Effects of habitat quality and ambient hyporheic flows on salmon spawning site selection

Understanding the role of stream hydrologic and morphologic variables on the selection of spawning sites by salmonid fishes at high resolution across broad scales is needed for effective habitat restoration and protection. Here we used remotely sensed meter-scale channel bathymetry for a 13.5 km reach of Chinook salmon spawning stream in central Idaho to describe habitat quality and set boundary conditions for a two-dimensional surface water model coupled with a three-dimensional hyporheic flux model. Metrics describing ambient hyporheic flow intensity and habitat quality, which is quantified as a function of stream hydraulics and morphology, were compared to the locations of nests built by female salmon. Nest locations were predicted most accurately by habitat quality followed by channel morphology (i.e., riffles location). As a lesser degree than habitat quality, water surface curvature was also a good indicator of spawning location because its intensity can identify riffle morphology. The ambient hyporheic flow predicted at meter-scale resolution was not a strong predictor of redd site selection. Furthermore, the study suggests direct morphological measurements obtained from easily measured channel bathymetry data could enable effective and rapid assessments of salmon spawning channels across broad areas.

Dam operations may improve aquatic habitat and offset negative effects of climate change

Dam operation impacts on stream hydraulics and ecological processes are well documented, but their effect depends on geographical regions and varies spatially and temporally. Many studies have quantified their effects on aquatic ecosystem based mostly on flow hydraulics overlooking stream water temperature and climatic conditions. Here, we used an integrated modeling framework, an ecohydraulics virtual watershed, that links catchment hydrology, hydraulics, stream water temperature and aquatic habitat models to test the hypothesis that reservoir management may help to mitigate some impacts caused by climate change on downstream flows and temperature. To address this hypothesis we applied the model to analyze the impact of reservoir operation (regulated flows) on Bull Trout, a cold water obligate salmonid, habitat, against unregulated flows for dry, average, and wet climatic conditions in the South Fork Boise River (SFBR), Idaho, USA.

Mapping river bathymetries: evaluating topobathymetric LiDAR survey: River bathymetry revealed

Advances in topobathymetric LiDARs could enable rapid surveys at sub-meter resolution over entire stream networks. This is the first step to improving our knowledge of riverine systems, both their morphology and role in ecosystems. The Experimental Advanced Airborne Research LiDAR B (EAARL-B) system is one such topobathymetric sensor, capable of mapping both terrestrial and aquatic systems. Whereas the original EAARL was developed to survey littoral areas, the new version, EAARL-B, was also designed for riverine systems but has yet to be tested. Thus, we evaluated the ability of EAARL-B to map bathymetry and floodplain topography at sub-meter resolution in a mid-size gravel-bed river. We coupled the EAARL-B survey with highly accurate field surveys (0.03m vertical accuracy and approximately 0.6 by 0.6m resolution) of three morphologically distinct reaches, approximately 200m long 15m wide, of the Lemhi River (Idaho, USA). Both point-to-point and raster-to-raster comparisons between ground and EAARL-B surveyed elevations show that differences (ground minus EAARL-B surveyed elevations) over the entire submerged topography are small (root mean square error, RMSE, and median absolute error, M, of 0.11m), and large differences (RMSE, between 0.15 and 0.38m and similar M) are mainly present in areas with abrupt elevation changes and covered by dense overhanging vegetation. RMSEs are as low as 0.03m over paved smooth surfaces, 0.07m in submerged, gradually varying topography, and as large as 0.24m along banks with and without dense, tall vegetation. EAARL-B performance is chiefly limited by point density in areas with strong elevation gradients and by LiDAR footprint size (0.2m) in areas with topographic features of similar size as the LiDAR footprint.