Peak-flow frequency analysis is essential for flood insurance studies, floodplain management, and the design of transportation infrastructure. In recent decades, better understanding of long-term hydroclimatic persistence, as well as concerns about potential climate change and land-use change have caused the stationarity assumption, underpinning for flood-frequency analysis, to be reexamined. The federal guidelines of Bulletin 17B (Interagency Advisory Committee on Water Data, 1982) and the recent updates in Bulletin 17C (England et al., 2018) recognize that the conventional assumptions for performing flood frequency analyses (e.g., the annual time series is a representative time sample of random homogeneous events and that the stochastic processes that generate floods are stationary or invariant in time) are violated in some cases. In the ‘Future Studies’ section of Bulletin 17C, the authors identify the need for additional studies related to “Guides for determining dynamic flood frequency curves that vary with time, incorporating climate indices, changing basin characteristics, and addressing potential nonstationary climate conditions.” To that end, the U.S. Geological Survey (USGS), in cooperation with the Federal Highway Administration (FHWA), has undertaken a multi-year national study to research methods for detecting and addressing potential trends in flood frequency distributions because of changes in climate, land use, land cover, and other potential drivers of change. The objectives of that study are to 1) identify where (and when) peak flow trends and change points have (or have not) occurred; 2) attribute causes of the trends; and 3) determine methods for modifying standard flood frequency analysis in regions where substantial change has (or is expected) to occur. The expected complexities and problems encountered on a national scale would be significantly informed by enhancing the national study with more focused regional studies. Illinois, Iowa, Minnesota, North Dakota, and South Dakota form a multi-state region in the Midwest with complex variability in topography, ecoregions, land use, land cover, and climate. The region is characterized by abrupt transitions among high-relief, semi-mountainous areas, well-drained moderate-elevation plains, poorly drained low-elevation glaciated prairies, and other complex geologic and hydrologic features. The land use is mainly pasture and small grains to the west. The Corn Belt covers much of the center and eastern portion of the region and has recently expanded to the west and north. The region also contains two major urban-suburban areas, centered on Chicago and Minneapolis-St. Paul, which have expanded substantially during the proposed study period and thus are another source of nonstationarity. Although many important water-management activities are independently conducted within the sociopolitical state boundaries, regional hydrologic complexities transcend the state boundaries and are more efficiently handled from a regional perspective. This proposed investigation will be conducted in direct collaboration with the ongoing national study and will focus on specific issues that have the most impact on flood-frequency analysis within the multi-state region in the Midwest. To investigate potential nonstationarity of peak flows in the multi-state region, preliminary exploratory peak-flow trend analyses were conducted. Figure 1 (in the Documents folder) shows significant (p<0.1) annual peak flow trends for all USGS streamgages with continuous record for 1960–2016; only trend directions (not magnitudes) are shown if significant trends are detected. There is strong spatial coherence in the historical trends for 1960–2016, with widely occurring downtrends for much of the conterminous U. S. and regional pockets of uptrends in the Northern Prairies (Eastern Dakotas/Western Minnesota), Upper Midwest and Central Great Lakes, and much of the New England and Mid-Atlantic states.

The overall goal of this study is to evaluate the combined effects of multidecadal climatic persistence (including hydroclimatic shifts), gradual climate change, and land-use change on peak-flow frequency analyses in the multi-state region in the Midwest. This study is intended to provide a framework for addressing potential nonstationarity issues in statewide flood-frequency updates that commonly are conducted by the USGS in cooperation with state DOTs throughout the nation on an ongoing basis. This will be achieved through the following primary objectives: 1. Define spatial and temporal characteristics of climatic persistence/change affecting annual peak flows in the multi-state region. 2. Develop and apply a statistical methodology for estimating changes in peak-flow frequency distributions in the multi-state region in relation to climatic persistence/change and urbanization; the effects of rural and land-use change will only be investigated in an exploratory manner. 3. Investigate methods for addressing regional climatic persistence/change and land-use change in peak-flow frequency analysis. To the extent possible, estimates of trend-adjusted flood magnitudes for various exceedance levels (such as the 10-percent or 1-percent annual exceedance probability) will be provided for comparison to previously published estimates.

The results of the proposed investigation will be presented in two peer-reviewed USGS Data Releases, two Scientific Investigations Reports (SIR; online only), two journal articles and a USGS Fact Sheet. Tasks include efforts to: (1) Publicly release watershed-based climate data (metrics of precipitation and temperature) on a monthly time scale and summed to annual seasonal and annual total values. These data will also contain trend results for trends in climate metrics in annual peak streamflow and climate variables (Years 1 and 2); (2) Characterize the effects of natural hydroclimatic shifts and potential climate change on annual peak flows in Midwest: Illinois, Iowa, Minnesota, North Dakota and South Dakota (Years 1 and 2); (3) Analyze the seasonality of flood peaks in the region and their trends and implications for trend attribution (Years 1 and 2); (4) Evaluate the effect of urbanization on flood-peaks in major metropolitan areas in the study region (Years 2 and 3); (5) Publicly release data that compares adjustment methods at individual sites, such as multiple flood-frequency analysis results (Years 3 and 4); (6) Summarize methods for addressing regional hydroclimatic shifts, climate change, and land-use change in peak-flow frequency analyses in the Midwest (Years 3 and 4); (7) Succinctly summarize the above work in a short fact sheet that provides links to project products to make it easy to find data and results. This summarization is intended to be handed out at future meetings, sent to managers for an overview, etc. (Year 4).

The funding level requested per state is $55,600 per year for four years. USGS will commit to additional cooperative funding of $30,000 per year for each participating state. Inclusion of additional states beyond the initial 5-state area can be accommodated, contingent on maintenance of a contiguous block of states and full funding commitments (additional state(s) and USGS). Minimum required funding is a full 4-year funding commitment from at least four contiguous states. Primary research efforts will be conducted by a team composed of USGS researchers from participating states. Collaborative research and/or funding from other Federal agencies also is encouraged.