This study combined flood-frequency analysis, 1-D (one-dimensional) hydraulic modeling using HEC-RAS, and flood-loss modeling using FEMA's Hazus-MH (Hazards U.S. Multi-Hazard) in order to: 1) quantify how different flood-frequency methodologies affect flood-risk assessments, and 2) quantify the impacts of different floodplain-management scenarios along the Middle Mississippi River (MMR). The nine scenarios tested here included various combinations of flood-frequency methodology, buyouts, and levee configurations. The levee configurations analyzed included: 1) current levee configuration, 2) no levees, 3) a 1500 m levee setback, 4) a 1000 m levee setback, and 5) a customized levee setback designed to maximize protection around existing infrastructure. Two study reaches were chosen: (1) an Urban Study Reach within St. Clair and northern Monroe Counties, IL, with levees designed to withstand the 500-year flood and (2) an Agricultural Study Reach within Union and Jackson Counties, IL, with <100-year levees. A flood-frequency analysis was completed for the St. Louis, MO gauging station, and detailed building inventories were used to estimate flood losses on a structure-by-structure basis (Hazus-MH UDF analysis) for an array of floods ranging from the 2- to the 500-year events. These flood-loss estimates were combined with a stochastic levee-failure model. Finally, estimated flood damages from Hazus-MH were integrated across the full range of flood recurrences in order to calculate expected annual damage (EAD). This study's flood-frequency analysis and corresponding flood-loss assessment demonstrate how differences in flood-frequency methodology can significantly impact flood-risk assessments. EAD based on the UMRSFFS (Upper Mississippi River System Flow Frequency Study) flood frequencies was 68% ($45.4 million) lower than EAD based on this study's flood frequencies. This decrease in EAD demonstrates that the UMRSFFS flood frequencies and corresponding stages may significantly underestimate flood risk within the Urban Study Reach. The 100-year discharge in the UMRSFFS appears to be underestimated by an estimated 17% (187,000 cfs), resulting in a 10% (1.6 m) underestimation of the 100-year flood level. Given the magnitude of the EAD, discharge, and stage differences documented here, a reanalysis of the MMR flood frequencies, flood profiles, and flood maps should be considered. The hydraulic modeling completed here showed that levee setbacks and levee removal successfully reduce stages for all recurrence intervals. For the 100-year flood, average reductions ranged from 0.20 m for a 1000 m levee setback to 1.61 m with levees removed. In general, stage reductions increased with increasing discharge and with increasing setback distance. The flood-level reductions are attributed to increased floodwater storage and conveyance across the reconnected floodplain. Compared to the current conditions, the levee setback and levee removal scenarios tested here reduced flood losses for large, infrequent flooding events but increased flood losses for smaller, more frequent flood events. When combined with buyouts of unprotected structures, levee setbacks reduced flood losses for all recurrence intervals. The 1000 m and 1500 m levee setbacks required buyouts in order to reduce EAD; however, a levee setback carefully planned around existing high-value structures reduced EAD with or without buyouts. The planned levee setback configuration combined with buyouts resulted in the largest decreases in EAD: a $16.8 million (55%) decrease in the Urban Study Reach and an $8.3 million (93%) decrease in the Agricultural Study Reach. Overall, this project showed that levee setbacks in combination with buyouts are an economically viable approach for flood-risk reduction along the study reaches and likely elsewhere where levees are widely employed for flood control. Designing a levee setback around existing high-value infrastructure can maximize the benefit of the setback while simultaneously minimizing the costs. Potentially, this type of planned levee configuration could be used as a template for the replacement of aging or failing levee systems.

During the late 19th and into the 20th Century, the Tisza River’s vast floodplain-wetland systems were largely disconnected by levees, facilitating "reclamation" for agriculture and resulting in an estimated loss of over 90% of historical wetlands. While levees have been successful in preventing catastrophic flooding for a century, Lower Tisza flood stages continue to rise partially due to aggradation and increased roughness on the confined floodplain. The decrease in the Tisza's current floodway carrying capacity has reduced the flood-protection level of the Tisza's aging levee system. Recently in Hungary, "Room for the River" policies have gained more prominence. For the first of three papers for this dissertation, I assessed eight potential floodplain-reconnection scenarios that would provide more room for the river between Csongrád, Hungary and the Hungary-Serbia border. A novel framework using hydrodynamic and geospatial modeling was used to perform planning-level evaluations of the tradeoffs between floodplain-reconnection scenarios and enhancement of the existing levee system. The scenarios evaluated include levee removal and levee setbacks to strategically reconnect historical wetlands while reducing flood levels. Scenario costs and human population impacts were also assessed. Impacts of reconnecting the Lower Tisza floodplain were compared to heightening levees, the prevailing strategy over the previous century. From a purely construction-cost perspective, heightening Lower Tisza levees is potentially the most cost-effective and politically expedient solution (i.e., impacts the least number of people). However, levee heightening does not solve the long-term problem of reduced flood conveyance, nor does it result in wetland reconnection or enhancement of other floodplain ecosystem services. The suite of reconnection options evaluated provides engineers, planners, and decision makers a framework from which they can further evaluate potential flood-risk reduction options. At least three of the eight reconnection scenarios (setting the western levee back, 1500-meter, and 2000-meter setbacks) along the Lower Tisza demonstrated that floodplain-wetland reconnection is possible while achieving the objectives of minimizing impacts on human populations and reducing flood heights. The Illinois River has a similar history to the Tisza. Levees were constructed, and wetlands were drained during the late 19th and early 20th Centuries. By the mid-1920’s, Illinois River levee systems became increasingly difficult for private landowners and the state to maintain as commodity prices fell and flood levels increased. However, the 1928 U.S. Flood Control act shifted a substantial portion of the burden of flood mitigation from local landowners to the federal government, preventing the dissolution of levee districts. While these levee systems have facilitated floodplain agricultural production and development for the last century, disconnecting the river from its floodplain has led to concerns about the negative impacts of levees on the physical and biological systems of the Illinois River Valley. Recent studies have emphasized approaches that would result in setting back or removing levees in order to naturalize portions of large river-floodplain systems, including the Illinois. The costs and benefits of such projects have shown potential restoration benefits may outweigh potential costs, but these studies have not demonstrated the specific levee districts which have the highest reconnection potential from an economic standpoint. The second paper for this dissertation used geospatial methods to fill this gap by assessing the National Commodity Crops Productivity Index (NCCPI) soil values and agricultural production and profit values for corn and soybeans in 32 individual levee districts along a 235-km segment of the Lower Illinois River. In general, soil productivity index values were lower for Illinois River levee districts compared to the county averages in which the districts are located. Over the five-year study period from 2010-2014, the total agricultural profits in the levee districts ranged from $18-61 million. Several levee districts have relatively low per hectare agricultural values when compared to wetland benefit studies, indicating these protected floodplain areas may be suitable for reconnection. For the third and final dissertation paper I used a novel hydrodynamic, geospatial, economic, and habitat suitability framework to assess the tradeoffs of strategically reconnecting the 125-km La Grange Segment (LGS) of the Lower Illinois River to its floodplain in order to decrease flood risk, improve floodplain habitats, and limit the costs of reconnection. Costs included building-associated losses, lost agricultural profits, and total levee removal and construction costs. Modeled scenarios demonstrated that while flood heights and environmental benefits are maximized through the most aggressive levee setbacks and removals, these scenarios also have the highest economic costs. However, the tradeoff of implementing lower-cost scenarios is that there would be less flood-height reduction and less floodplain habitat available. Several levee districts had high potential for reconnection based on limiting potential damages as well as providing suitable floodplain habitat. To implement large-scale strategic floodplain reconnection along the LGS, opportunity costs ranged from $1.1-$4.3 billion. As such, payments for ecosystem services will likely be necessary to compensate landowners for building losses and decreased long-term agricultural production that result in an overall flood-reduction benefit, increased floodplain wetlands, and most-soil plant habitat.

When analyzing the flood risk using multi-criteria analysis it is necessary to estimate