No abstract

In this article, the mean daily streamflow at 139 streamflow-gaging stations (sites) in the southern and southeastern United States are analyzed for spatial and temporal patterns. One hundred and thirty-nine individual time-series of mean daily streamflow were reduced to five aggregated time series of Z scores for clusters of sites with similar temporal variability. These aggregated time-series correlated significantly with a time-series of several climate indices for the period 1950–2015. The mean daily streamflow data were subset into six time periods—starting in 1950, 1960, 1970, 1980, 1990, and 2000, and each ending in 2015, to determine how streamflow trends at individual sites acted over time. During the period 1950–2015, mean monthly and seasonal streamflow decreased at many sites based on results from traditional Mann–Kendall trend analyses, as well as results from a new analysis (Quantile-Kendall) that summarizes trends across the full range of streamflows. A trend departure index used to compare results from non-reference with reference sites identified that streamflow trends at 88% of the study sites have been influenced by non-climatic factors (such as land- and water-management practices) and that the majority of these sites were located in Texas, Louisiana, and Georgia. Analysis of the results found that for sites throughout the study area that were influenced primarily by climate rather than human activities, the step increase in streamflow in 1970 documented in previous studies was offset by subsequent monotonic decreases in streamflow between 1970 and 2015.

Aim Stream fish distributions are hypothesized to be strongly associated with landscape characteristics at multiple scales. Variation in flow regimes and intensity of landscape disturbance are associated with stream fish distributions; however, relationships are poorly understood in many high‐diversity regions. Our objective was to identify occurrence relationships between fish distributions and streamflow and landscape characteristics in the south‐central United States. Location Our study area was the central Red River catchment in Oklahoma, Texas and Arkansas, USA. Methods We used existing fish surveys to model the occurrence of a diverse, warmwater assemblage among hydraulic response units (HRUs). We used multispecies occupancy modelling to identify variation in occurrence probability among 111 stream fishes in relation to landscape disturbance and flow regime characteristics. Results We found occurrence relationships with landscape disturbance and 11 metrics comprising all flow‐regime components. The relationships varied within both major species groups and some genera. Frequency and duration were the most common metrics underlying flow regime relationships. More common stream fishes tended to be positively associated with higher levels of landscape disturbance and flow regime metrics representing variability; conversely, narrow‐ranged fishes tended to be negatively associated. Occurrence relationships with flow metrics representing high‐flow events were predominately negative. As expected, many species were strongly associated with ecoregion with landscape disturbance and flow relationships held constant. Main conclusions Our study informs land use and water management decisions and stream fish conservation at multiple spatial scales. Collectively, the findings suggest potential homogenization of the Red River fish assemblage with increased landscape disturbance and streamflow variability. A reduction in landscape disturbance and maintenance of natural flow patterns at coarser scales may benefit endemic and narrow‐ranged fishes. Our findings also help guide finer‐scale land use and water management decisions by identifying stream network areas with a high occurrence probability of less tolerant fishes.

No abstract

The Precipitation-Runoff Modeling System (PRMS) was used to develop and calibrate a streamflow and water balance model for the Red River Basin as part of the U.S. Geological Survey National Water Census, a research effort focused on developing innovative water accounting tools and conducting assessments of water use and availability at regional and national spatial scales. The PRMS is a deterministic model that simulates the effects of climate, land cover, and water use on watershed hydrology on the basis of physical processes and spatial attributes of the watershed. The model was used to estimate streamflow at daily and monthly temporal scales for the 1980-2016 period and to evaluate the impacts of natural and anthropogenic influences on streamflow and water budget components.Sixty-seven percent of streamgages were calibrated successfully for the monthly time step and 43 percent of streamgages were successfully calibrated for the daily time step. Some of the challenges of calibrating streamgages included estimating low amounts of streamflow in dry areas of the basin and accurately representing watershed characteristics related to evapotranspiration in the basin, among other factors. The model estimated streamflow with some accuracy for 36 percent and 26 percent of the 73 streamgages used to evaluate the model at monthly and daily time steps, respectively. Relative to no-water-use conditions, water use increased streamflow volumes (that is, return flow from reservoir releases) the most on the main stem of the Red River, the North Fork of the Red River, and the Ouachita River. Water withdrawal decreased streamflow volumes most in the Red River near the outlet of the basin and in Caney Creek. Streamflow volumes on the North Fork of the Red River changed most as a result of water use. The Red River Basin PRMS model provided estimates of streamflow that were limited in their accuracy by (1) the availability of accurate water-use data; (2) the coarse resolution of spatial parameters (such as those for impervious area or plant canopy), which leads to the homogenization of physical features in small watersheds in the model domain; and (3) the accuracy of spatial patterns of precipitation distribution across the model domain. Improvements in the quality and quantity of available water-use data and finer resolution spatial parameter and climate data could lead to the development of better-informed models in the future that are capable of making more accurate estimates of streamflow, because they are more representative of physical and hydrologic conditions in the Red River Basin.