To improve flood-frequency estimates at rural streams in Mississippi, annual exceedance probability (AEP) flows at gaged streams in Mississippi and regional-regression equations, used to estimate annual exceedance probability flows for ungaged streams in Mississippi, were developed by using current geospatial data, additional statistical methods, and annual peak-flow data through the 2013 water year. The regional-regression equations were derived from statistical analyses of peak-flow data, basin characteristics associated with 281 streamgages, the generalized skew from Bulletin 17B (Interagency Advisory Committee on Water Data, 1982), and a newly developed study-specific skew for select four-digit hydrologic unit code (HUC4) watersheds in Mississippi. Four flood regions were identified based on residuals from the regional-regression analyses. No analysis was conducted for streams in the Mississippi Alluvial Plain flood region because of a lack of long-term streamflow data and poorly defined basin characteristics. Flood regions containing sites with similar basin and climatic characteristics yielded better regional-regression equations with lower error percentages. The generalized least squares method was used to develop the final regression models for each flood region for annual exceedance probability flows. The peak-flow statistics were estimated by fitting a log-Pearson type III distribution to records of annual peak flows and then applying two additional statistical methods: (1) the expected moments algorithm to help describe uncertainty in annual peak flows and to better represent missing and historical record; and (2) the generalized multiple Grubbs-Beck test to screen out potentially influential low outliers and to better fit the upper end of the peak-flow distribution. Standard errors of prediction of the generalized least-squares models ranged from 28 to 46 percent. Pseudo coefficients of determination of the models ranged from 91 to 96 percent. Flood Region A, located in north-central Mississippi, contained 27 streamgages with drainage areas that ranged from 1.41 to 612 square miles. The 1% annual exceedance probability had a standard error of prediction of 31 percent which was lower than the prediction errors in Flood Regions B and C.

To improve flood-frequency estimates at rural streams in Mississippi, annual exceedance probability (AEP) flows at gaged streams in Mississippi and regional-regression equations, used to estimate annual exceedance probability flows for ungaged streams in Mississippi, were developed by using current geospatial data, additional statistical methods, and annual peak-flow data through the 2013 water year. The regional-regression equations were derived from statistical analyses of peak-flow data, basin characteristics associated with 281 streamgages, the generalized skew from Bulletin 17B (Interagency Advisory Committee on Water Data, 1982), and a newly developed study-specific skew for select four-digit hydrologic unit code (HUC4) watersheds in Mississippi. Four flood regions were identified based on residuals from the regional-regression analyses. No analysis was conducted for streams in the Mississippi Alluvial Plain flood region because of a lack of long-term streamflow data and poorly defined basin characteristics. Flood regions containing sites with similar basin and climatic characteristics yielded better regional-regression equations with lower error percentages. The generalized least squares method was used to develop the final regression models for each flood region for annual exceedance probability flows. The peak-flow statistics were estimated by fitting a log-Pearson type III distribution to records of annual peak flows and then applying two additional statistical methods: (1) the expected moments algorithm to help describe uncertainty in annual peak flows and to better represent missing and historical record; and (2) the generalized multiple Grubbs-Beck test to screen out potentially influential low outliers and to better fit the upper end of the peak-flow distribution. Standard errors of prediction of the generalized least-squares models ranged from 28 to 46 percent. Pseudo coefficients of determination of the models ranged from 91 to 96 percent. Flood Region C, located in the southwest corner of Mississippi, contained 120 streamgages with drainage areas that ranged from 0.05 to 1,010 square miles. The 1% annual exceedance probability had a standard error of prediction of 41 percent.

To improve flood-frequency estimates at rural streams in Mississippi, annual exceedance probability (AEP) flows at gaged streams in Mississippi and regional-regression equations, used to estimate annual exceedance probability flows for ungaged streams in Mississippi, were developed by using current geospatial data, additional statistical methods, and annual peak-flow data through the 2013 water year. The regional-regression equations were derived from statistical analyses of peak-flow data, basin characteristics associated with 281 streamgages, the generalized skew from Bulletin 17B (Interagency Advisory Committee on Water Data, 1982), and a newly developed study-specific skew for select four-digit hydrologic unit code (HUC4) watersheds in Mississippi. Four flood regions were identified based on residuals from the regional-regression analyses. No analysis was conducted for streams in the Mississippi Alluvial Plain flood region because of a lack of long-term streamflow data and poorly defined basin characteristics. Flood regions containing sites with similar basin and climatic characteristics yielded better regional-regression equations with lower error percentages. The generalized least squares method was used to develop the final regression models for each flood region for annual exceedance probability flows. The peak-flow statistics were estimated by fitting a log-Pearson type III distribution to records of annual peak flows and then applying two additional statistical methods: (1) the expected moments algorithm to help describe uncertainty in annual peak flows and to better represent missing and historical record; and (2) the generalized multiple Grubbs-Beck test to screen out potentially influential low outliers and to better fit the upper end of the peak-flow distribution. Standard errors of prediction of the generalized least-squares models ranged from 28 to 46 percent. Pseudo coefficients of determination of the models ranged from 91 to 96 percent. Flood Region B, located along the eastern border of Mississippi, contained 134 streamgages with drainage areas that ranged from 0.15 to 1,750 square miles. The 1% annual exceedance probability had a standard error of prediction of 35 percent.

A combination of discrete and daily-aligned groundwater levels for the Mississippi River Valley alluvial aquifer clipped to the Mississippi Alluvial Plain, as defined by Painter and Westerman (2018), with corresponding metadata are based on processing of U.S. Geological Survey National Water Information System (NWIS) (U.S. Geological Survey, 2020) data. The processing was made after retrieval using aggregation and filtering through the infoGW2visGWDB software (Asquith and Seanor, 2019). The nomenclature GWmaster mimics that of the output from infoGW2visGWDB. Two separate data retrievals for NWIS were made. First, the discrete data were retrieved, and second, continuous records from recorder sites with daily-mean or other daily statistics codes were retrieved. Each dataset was separately passed through the infoGW2visGWDB software to create a "GWmaster discrete" and "GWmaster continuous" and these tables were combined and then sorted on the site identifier and date to form the data products described herein. A sweep through the combined dataset (the "database") was made to isolate duplicate observations, or observations for the same well and on the same day. If a discrete value was present, it was retained as authoritative for the day and in descending order of priority daily-mean, daily-maximum, and daily minimum. Therefore, only a single record for a well and day are present in the dataset. The duplicate search removed 876 records and 31 wells were involved; in total, this is about 0.3 percent of the database. References: Asquith, W.H., Seanor, R.C., 2019, infoGW2visGWDB—An R groundwater data-processing utility for manipulating, checking the veracity, and converting an "infoGW" object to the "GWmaster" object for the visGWDB software with demonstration for the Mississippi River Valley alluvial aquifer: U.S. Geological Survey software release, Reston, Va., https://doi.org/10.5066/P9MK0B6L. Painter, J.A., and Westerman, D.A., 2018. Mississippi Alluvial Plain extent, November 2017: U.S. Geological Survey data release, https://doi.org/10.5066/F70R9NMJ. U.S. Geological Survey, 2020, USGS water data for the Nation: U.S. Geological Survey National Water Information System database, accessed April 2, 2020, at https://doi.org/10.5066/F7P55KJN.

The Mississippi Alluvial Plain (MAP) has become one of the most important agricultural regions in the US, and it relies heavily on a groundwater system that is poorly understood and shows signs of substantial change. The heavy use of the available groundwater resources has resulted in significant groundwater-level declines and reductions in base flow in streams within the MAP. These impacts are limiting well production and threatening future water-availability for the region. This product will help not only scientists in our center, but also at a national level. This product will also be part of a larger study encompassing the Mississippi Alluvial Plain region. The Mississippi Alluvial Plain extent was delineated using GIS tools to represent the geographic extent of the Mississippi Alluvial Aquifer through incorporation of elevation information, geomorphology knowledge, ecological region extent, and previously published extents for part of the MAP region. The current MAP extent represents version 1.0. Future changes to the MAP extent will be tracked through increasing version numbers.

Groundwater is a vital resource to the Mississippi embayment region of the central United States. Regional and integrated assessments of water availability that link physical flow models and water quality in principal aquifer systems provide context for the long-term availability of these water resources. An innovative approach using machine learning was employed to predict groundwater pH across drinking water aquifers of the Mississippi embayment. The region includes two principal regional aquifer systems; the Mississippi River Valley alluvial (MRVA) aquifer and the Mississippi embayment aquifer system that includes several regional aquifers and confining units. Based on the distribution of groundwater use for drinking water, the modeling effort was focused on the MRVA, Middle Claiborne aquifer (MCAQ), and Lower Claiborne aquifer (LCAQ)of the Mississippi embayment aquifer system. Boosted regression tree (BRT) models (Elith and others, 2008; Kuhn and Johnson, 2013) were used to predict pH to 1-km raster grid cells of the National Hydrologic Grid (Clark and others, 2018). Predictions were made for 7 aquifer layers (1 MRVA, 4 MCAQ, 2 LCAQ) following the hydrogeologic framework used in a regional groundwater flow model (Hart and others, 2008). Explanatory variables for the BRT models included attributes associated with well position and construction, surficial variables, and variables extracted from a MODFLOW groundwater flow model for the MISE (Haugh and others, 2020a,b). For a full description of modeling workflow see Knierim and others (2020).

This data set provides estimated and measured streamflow data and hydrograph-separation results for five sites located in northwest Mississippi. Streamflow data were collected by the U.S. Geological Survey (USGS) and the U.S. Army Corps of Engineers. Hydrograph-separation results provide runoff and baseflow estimates at each site that were calculated using four methods: PART, HYSEP Fixed, HYSEP Local Minimum, and BFI Standard, as well as an average base flow index (BFI) for all four methods.

To identify the degree of hydrologic alteration of streams in the Mississippi Alluvial Plain (MAP), we used random forest (RF) regression methods (Breiman, 2001) to model the relation between six selected streamflow characteristics and explanatory variables (such as drainage area, precipitation, soils, and other watershed characteristics). RFs were chosen for this study because they have been proven to be more robust and accurate than traditional linear regression methods (Carlisle and others, 2010; Lawler and others, 2006; Prasad and others, 2006; Cutler and others, 2007). Estimated expected monthly mean streamflow from the RF models were compared to observed monthly mean streamflow at 68 sites located within the MAP and two adjacent Level III Ecoregions. We also used an additional eight sites to compare estimated expected streamflow, generated by the RF models, and observed streamflow for characteristics of flood frequency, high streamflow duration, number of zero streamflow days, frequency of low-pulse spells, and high streamflow discharge. This data release includes the explanatory variables (input data) used in the random-forest models (Breiman, 2001) to determine expected flows (output data) at 76 sites in the MAP. The geospatial dataset contains the point and watershed features for the sites used in the analyses. These data were used to support the findings in the journal article titled "Quantifying Hydrologic Alteration in an Area Lacking Current Reference Conditions—The Mississippi Alluvial Plain of the South-Central U.S." by Hart and Breaker (2018). References: Breiman, L. 2001, Random forests: Machine Learning, v. 45, p.5–32. Carlisle, D.M., Falcone, J., Wolock, D.M., Meador, M.R., and Norris, R.H., 2010, Predicting the natural streamflow regime: models for assessing hydrological alteration in streams: River Research and Applications, v. 26, p.118–136. Cutler, D.R., Edwards, T.C. Jr, Beard, K.H., Cutler, A., Hess, K.T., Gibson, J., Lawler, J.J., 2007, Random forests for classification in ecology: Ecology v. 88, p.2783–2792. Lawler, J.J., White, D., Neilson, R.P., and Blaustein, A.R., 2006, Predicting climate-induced range shifts: model differences and model reliability: Global Climate Change Biology v. 12, p.1568–1584. Prasad, A.M., Iverson, L.R., Liaw, A., 2006, Newer classification and regression tree techniques: bagging and random forests for ecological prediction: Ecosystems v. 9, p.181–199.