Stakeholder-led science : engaging resource managers to identify science needs for long-term management of floodplain conservation lands

Floodplains pose challenges to managers of conservation lands because of constantly changing interactions with their rivers. Although scientific knowledge and understanding of the dynamics and drivers of river-floodplain systems can provide guidance to floodplain managers, the scientific process often occurs in isolation from management. Further, communication barriers between scientists and managers can be obstacles to appropriate application of scientific knowledge. With the coproduction of science in mind, our objectives were the following: (1) to document management priorities of floodplain conservation lands, and (2) identify science needs required to better manage the identified management priorities under nonstationary conditions, i.e., climate change, through stakeholder queries and interactions. We conducted an online survey with 80 resource managers of floodplain conservation lands along the Upper and Middle Mississippi River and Lower Missouri River, USA, to evaluate management priority, management intensity, and available scientific information for management objectives and conservation targets. Management objectives with the least information available relative to priority included controlling invasive species, maintaining respectful relationships with neighbors, and managing native, nongame species. Conservation targets with the least information available to manage relative to management priority included pollinators, marsh birds, reptiles, and shore birds. A follow-up workshop and survey focused on clarifying science needs to achieve management objectives under nonstationary conditions. Managers agreed that metrics of inundation, including depth and extent of inundation, and frequency, duration, and timing of inundation would be the most useful metrics for management of floodplain conservation lands with multiple objectives. This assessment provides guidance for developing relevant and accessible science products to inform management of highly dynamic floodplain environments. Although the problems facing managers of these lands are complex, products focused on a small suite of inundation metrics were determined to be the most useful to guide the decision making process.


INTRODUCTION
Floodplains owe their high biodiversity and productivity to dynamic spatial and temporal interactions with the adjacent river (Bayley 1995, Tockner et al. 2000, Tockner and Stanford 2002).Floodplains provide productive soils, suitable topography, and abundant water resources that have historically driven agricultural, urban, and industrial development on these lands.The widespread construction of levees to protect human uses from floods have reduced the frequency of inundation of floodplains, yet contributed to increased river stages during flood events and increased discharges downstream of leveed areas (Di Baldassarre et al. 2009, Remo et al. 2009, Heine and Pinter 2012).Recent, extreme floods on the large rivers have prompted reconsideration of the role of floodplains in regulating flooding processes, mitigating flood damages, and providing conservation values (Sparks 1995).As a result, large tracts of floodplains have been reconnected in large rivers in Europe (Buijse et al. 2002, Pahl-Wostl 2006, Hein et al. 2016) and the United States (U.S. Fish and Wildlife Service 2014).Within the state of Missouri USA, for example, approximately 35,000 hectares of floodplain lands have been acquired by state and federal agencies along the Missouri and Mississippi rivers, and additional lands are under various conservation easements (USGS, unpublished data).The majority of these properties have been converted from agricultural production to natural land cover and managed for conservation.
Management strategies of floodplain conservation lands range conceptually from active to passive (Galat et al. 1998).Actively managed floodplain lands typically rely on infrastructure, such as pumps, gates and constructed wetlands, for manipulating water levels on the floodplain (Fredrickson and Taylor 1982).Conversely, passively managed floodplains rely on riverfloodplain connectivity during high river stages for managing inundations (Galat et al. 1998).Floodplain management objectives and conservation targets may differ with the type of management strategy, but are generally met with a series of both short-and long-term management decisions.Short-term management decisions (seasonal to annual), such as whether to pump water to or from the property, are commonly more adaptable than long-term management decisions and dictated by seasonal hydrologic and climatic conditions.Long-term management decisions (multiyear to centennial), such as the type of forest to restore, often rely on an understanding of the range, variability, and projections of hydroclimatic conditions, that are derived from observed historical data (Stanturf et al. 2000).
Floodplain management is inherently difficult because of constantly changing interactions of floodplains with their rivers, wide ranges of variability, and anthropogenic modifications of in-channel conditions and throughout contributing drainage areas (Adams andPerrow 1999, Hughes et al. 2005).Challenges of managing floodplains are compounded when hydroclimatic stationarity cannot be assumed, and changing climate, land use, and/or water use combine to alter the magnitude, duration, and timing of hydrologic events (Olsen 2006).Relying on past hydrologic records to guide management practices is problematic under nonstationary conditions, because the historical conditions that have driven a system to its present state may not be the same conditions that will drive the system in the future (Milly et al. 2008).For example, dam construction has significantly altered hydrologic regimes worldwide over the past century, whereas climate change may drive future changes in hydrologic regimes.Therefore, problems may arise because the infrastructure and accompanying land uses designed for specific hydroclimatic conditions from the period of record may not support future hydroclimatic conditions.Incongruities between design specifications and driving conditions will lead to untenable land use decisions, ultimately leading to possible project failure and wasted resources.Analysis of trade-offs between different management actions across a range of future conditions may aid in identifying flexible and robust management actions (Poff et al. 2016, Singh et al. 2015).
Increased scientific knowledge of nonstationarity in riverfloodplain systems can provide guidance to floodplain managers; however, the scientific process, although well intentioned, often occurs in isolation from management.Scientific questions and results are often relevant to management, but rarely match management needs.Further, communication barriers between scientists and managers is a commonly cited obstacle to application of scientific knowledge (Wright 2007, Kocher et al. 2012).Coproduction of scientific knowledge, where research questions arise from interactions between researchers and information users, has led to successful use of that scientific knowledge in incorporating climate science in forest management (Lemos and Morehouse 2005, Dilling and Lemos 2011, Littell et al. 2012).With the coproduction of science in mind, our objectives were to (1) document management priorities of floodplain conservation lands, and (2) identify science needs required to better manage the identified management priorities under nonstationary conditions, i.e., climate change.Our approach was to determine management priorities, science needs, and constraints on using scientific information through stakeholder queries and interactions and to evaluate if priorities and science needs were consistent across the study area.We hypothesized that aspects of hydrology would emerge as common science needs across management priorities, given hydrology is a key driver of floodplain ecosystems.We highlight the collaborative process we used to identify science needs and discuss how these results provide guidance for developing relevant and accessible science to inform management of highly dynamic floodplain environments that often have multiple management objectives.

Study area
The project focused on floodplain conservation lands along the Upper Mississippi River (UMR), the Middle Mississippi River (MMR), and the Lower Missouri River (LMOR), USA (Fig. 1).The selected major river sections present a wide range of both natural and human-induced hydrologic and geomorphic variation and, as such, provide results that may be transferable to other large rivers (Table 1).The UMR (1060 kilometers) flows from the headwaters to the confluence with the Missouri River and is impounded with a series of 29 navigation locks and dams that create pools with relatively low variability in stage relative to the other two study sections (Schramm et al. 2015).The stage of the navigation pools can be managed to a limited extent by the U.S. Army Corps of Engineers (USACE) to influence connectivity to floodplains, but ultimately the regulation structures on the UMR have artificially maintained high water levels relative to historical conditions (Theiling and Nestler 2010).The LMOR is a 1300 kilometer (km) reach downstream of Gavin's Point Dam to the confluence with the Mississippi River near St. Louis, Missouri.The river is unchannelized in the 94 km between Gavin's Point Dam and Ponca, Nebraska, yet floodplain connectivity is compromised by channel incision (Jacobson and Galat 2006).Downstream of Ponca, Nebraska, river training structures (wing dikes) have confined the historically wide, shallow, and braided channel to a narrow and deep single thread channel, effectively creating a self-scouring channel (Jacobson et al. 2015).The river is largely disconnected from the adjacent floodplain through an extensive series of levees and channel incision.Connectivity between the main channel and the floodplain generally increases in the downstream direction along the LMOR.The LMOR is downstream from the largest reservoir system in North America with 91 km³ of total storage (U.S. Army Corps of Engineers 2006).The mainstem reservoir system is http://www.ecologyandsociety.org/vol21/iss3/art12/managed by the USACE for multiple purposes; management for floodplain connectivity, however, has been considered contradictory to the USACE flood-control mandate (Jacobson and Galat 2008).The hydrologic signal from scheduled reservoir operations and releases is attenuated, but remains discernible throughout the full length (Galat and Lipkin 2000).
The MMR is defined as 305 kilometers of the Mississippi River from the confluence of the Missouri River to the confluence with the Ohio River.Similar to the LMOR, the MMR lacks navigation dams and is channelized through river training structures.Levees constrain floodplain connectivity and preclude the lower return interval floods from accessing much of the floodplain in this section, much of which is used for agriculture (Remo et al. 2009).
The MMR experiences higher variability in flow conditions relative to the UMR because (1) reservoir storage is relatively low throughout the UMR drainage basin, including tributaries, and (2) dam operation signals from the UMR and LMOR have effectively attenuated upon reaching the MMR because of cumulative additions from tributaries.

Participants
We developed a three-step interactive approach to document management priorities and science needs of floodplain conservation lands (Fig. 2).Our target stakeholders were natural resource managers of floodplain conservation lands owned or managed by federal and state natural resource agencies (Appendix 1); these agencies provide public access to lands for recreation and hunting, as well as managing lands for ecosystem benefits.Governance structure of these properties differs by agency; therefore, our list includes personnel with a wide range of expertise and roles in the management of floodplain properties.

Data collection
We designed and distributed an online survey (Appendix 2) via email to all identified resource managers in the fall of 2014 and asked them to prioritize and rate information available to achieve selected management objectives and conservation targets (Table 2).This initial survey and selection of objectives and targets were developed with input from knowledgeable academic, state, and federal scientists and resource managers in addition to publicly available management plans.Objectives reflect broader management goals while conservation targets reflect species or biotic communities that are often used as measures of conservation success.For our purposes, conservation targets reflect both species and communities in need of conservation (e.g., threatened species, game species) and control (e.g., invasive species).Additional questions focused on the importance of selected sources of scientific information, and constraints on obtaining and using scientific information.Structured survey questions with scaled responses (i.e., low, medium, high priority) allowed for quantitative analysis of survey results.

Fig. 2.
Our objective to identify management priorities and science needs for long-term management of large-river floodplain conservation lands were met with a three-step process of interacting with floodplain managers through online surveys and a workshop.

Data analysis
We aggregated individual responses from the initial survey by river reach to test the effect of reach (categorical) and objective/target on priority score, i.e., low (1), medium (2), high (3), using twoway analysis of variance (alpha = 0.05).Because of consistent prioritization across reaches, differences in mean priority scores between objectives, aggregated across river reaches, were tested using Tukey's honest significant difference (HSD).Priority scores and information availability scores had different response scales, therefore they were standardized, each separately for management objectives and conservation targets.The standardized information value, I S , was subtracted from the standardized priority value, P S , to assess the information available to manage an objective or target relative to its priority, IP S .
Where, I i refers to information value for objective i, I min refers to the minimum information value among all objectives, I max refers to the maximum information value among all objectives, P i refers to priority score for objective i, P min refers to the minimum priority score among all objectives, and P max refers to the maximum priority score among all objectives.http://www.ecologyandsociety.org/vol21/iss3/art12/High values of IP S suggest that managers believe there is sufficient information available to manage an objective or target relative to its management priority, whereas low values suggest limited information relative to priority.
The second survey had fewer respondents, therefore we aggregated individual responses across reaches.We tested for differences in frequency of use and information value of different types of scientific information in addition to differences in information value and need for different inundation metrics using one-way analysis of variance.All statistical analyses were performed using computing environment R (R Core Team 2014).

Initial survey
We identified 80 individuals managing 155 floodplain conservation lands, totaling approximately 2400 km².Our initial survey had a response rate of 68%, with 54 individuals completing the survey.Eight individuals declined the survey invitation because of new responsibilities or inexperience and were removed from the list of floodplain managers.The respondents managed a total of 125 properties (80% of floodplain conservation lands identified).Approximately 81% of floodplain conservation lands along the Upper Mississippi River were represented (51/63), 67% of Middle Mississippi River properties (8/12), and 80% of Lower Missouri River properties (66/83).Seventeen percent (9/54) of the respondents represented federal agencies, 81% (44/54) represented state agencies, and 2% (1/54) represented a nonprofit organization.
There was an effect of management objective (F 16, 967 = 29.27,p < 0.001) and river reach (F 2, 967 = 6.55, p = 0.002) on priority scores, but no interaction effect between objective and river reach (F 32, 935 = 0.94, p = 0.56).Objectives with the highest priority scores included maintaining good relationships with neighboring land owners, providing public recreation opportunities, controlling invasive species, managing native nongame species, and managing endangered and threatened species (Tukey's HSD p < 0.05; Fig. 3A).Objectives with the lowest priority ranking included agricultural production, contaminant retention, and reducing flood risk.LMOR consistently had higher priority scores across all objectives in comparison to the UMR and MMR (Tukey's HSD p < 0.05), but the lack of an interaction effect suggests objective prioritization was ranked similarly across river reaches.Additional management objectives were identified by survey participants and included determining and meeting objectives of private landowners, timber production, birding, pecan production, education and interpretation, identifying alterations from historical conditions, linking private and public land conservation, oak savanna restoration, and reducing sediment deposition.
There was an effect of management objective (F 16, 859 = 11.21,p < 0.001) and river reach (F 2, 859 = 9.63, p < 0.001) on the perceived availability of information, but no interaction effect (F 32, 827 = 0.67, p = 0.92).Sufficient information was thought to be available to manage for agricultural production, game species, recreational opportunities, and wetlands while objectives with the least information available to manage include contaminant retention, nutrient retention, and restoration of bottomland prairies (Tukey's HSD p < 0.05; Figure 3A).More information was perceived to be available in the LMOR and UMR compared to the MMR (Tukey's HSD p < 0.05), but no interaction effect suggests objectives were ranked in terms of information availability similarly across river reaches.Standardized information availability scores relative to standardized priority scores reveal sufficient information relative to priority for agricultural production, managing game species, flood risk reduction, maintaining current conditions, and restoring wetlands, and limited information relative to priority for all other objectives (Fig. 4A).Objectives with the least information available relative to priority include controlling invasive species, maintaining respectful relationships with neighbors, and managing native, nongame species.Priority scores differed by conservation target (F 19, 1118 = 18.18, p < 0.001) and river reach (F 2, 1118 = 30.72,p < 0.001).Priority scores of conservation targets were consistent among river reaches (F 38, http://www.ecologyandsociety.org/vol21/iss3/art12/reach.Upon closer examination, Ring-necked Pheasant (Phasianus colchicus) and Piping Plover (Charadrius melodus) had higher priority in the LMOR compared with other reaches, and pallid sturgeon (Scaphirhynchus albus), Least Tern (Sternula antillarum), and Northern Bobwhite (Colinus virginianus) had higher priority in the LMOR and MMR compared to the UMR.Targets with the highest priority across reaches included waterfowl, invasive plants, song birds, and Bald Eagles (Haliaeetus leucocephalus; Tukey's HSD p < 0.05; Fig. 3B).The LMOR had higher priority scores across targets than MMR and UMR (Tukey's HSD p < 0.05).Additional conservation targets identified by participants included willows, bottomland hardwoods, fish spawning areas, swamp white oak (Quercus bicolor), silver maple (Acer saccharinum), furbearing mammals, aquatic vegetation, mast-producing floodplain forest species, endangered plant species, and Mourning Doves (Zenaida macroura).Perceived information availability differed by conservation targets (F 19, 967 = 15.63,p < 0.001) and river reach (F 2, 967 = 2.44, p = 0.03), but perceived information availability for conservation targets was consistent among river reaches (F 38,929 = 0.68, p = 0.93).Sufficient information is available to manage for white-tailed deer, Wild Turkey (Meleagris gallopavo), Ring-necked Pheasant, and waterfowl, but limited information is available to manage for pollinators, aquatic invasive species, reptiles, amphibians, and marsh birds (Tukey's HSD p < 0.05; Figure 3B).Perceived information availability was generally greater in the LMOR compared to MMR, with neither being different from the UMR (Tukey's HSD p < 0.05).Ring-necked pheasant, Northern Bobwhite (Colinus virginianus), game fish, Wild Turkey, and white-tailed deer had the greatest perceived information available relative to management priority whereas pollinators, marsh birds, reptiles, song birds, aquatic invertebrates, and amphibians had the least perceived information available relative to management priority (Fig. 4B).
Information from more experienced within-agency personnel was a significantly more important source of scientific information than word of mouth, peer-reviewed literature, white papers, and internal datasets (F 7,448 = 3.951, p < 0.001; Tukey's HSD p < 0.05); however, all sources of information were rated between important and very important.Time, funding, and disconnect between science and management activities were significantly more constraining to managers on their ability to obtain or use new scientific information in management decisions than credibility of scientific information, credibility of scientific process, and access to scientific journals (F 9,560 = 15.57,p < 0.001; Table 3).
Table 3. Mean value and standard deviation from 57 survey respondents for how limiting (1 = not limiting, 4 = very limiting) different constraints are when incorporating new scientific information into the decision-making process.An analysis of variance and Tukey's HSD test was performed on each dataset to determine statistical differences in mean rank between metrics.Superscript letters that differ indicate ranks that differ at alpha = 0.05.We anticipated managers would select objectives or conservation targets with little information available relative to management priority for which to develop conceptual models.However, it became clear through our discussions that floodplain conservation lands were generally managed for multiple objectives and selecting a single objective or target was not an efficient pathway to determine overall science needs to manage these properties.There was consensus among managers that http://www.ecologyandsociety.org/vol21/iss3/art12/primary abiotic factors, particularly metrics of floodplain inundation, were the variables of greatest interest to understand and predict as they drive secondary abiotic factors and define habitat for biota.However, managers were less concerned about needing models for biological endpoints linked to the abiotic factors, primarily because they were more comfortable identifying the biological link themselves for the biotic endpoint of interest.

Follow-up survey
Twenty-one floodplain managers participated in the follow-up survey to the workshop.Of these participants, 13 (62%) managed properties on the LMOR, four (19%) on the UMR, 3 (14%) on the MMR, and 1 (5%) on the Illinois River.Thirteen managers (62%) self-associated with an active management of their lands while the remaining 8 (38%) self-associated with passive management.
There were significant differences in the frequency of use (F 3,80 = 8.628, p < 0.001) of scientific information identified in the workshop as commonly used for management of floodplain conservation lands.Topography and elevation were used significantly more than hydrogeomorphic classifications of habitat or metrics on inundation, although this is likely due to limited availability of the latter two types of information (Tukey's HSD p < 0.05).The informational value attributed to types of information did not differ, with all types of information identified as informative to planning and management decisions (F 3,65 = 2.217, p = 0.100).Additional types of information identified by respondents as commonly used in management decisions included species occurrence from monitoring or existing databases, known wildlife needs, accessibility, location of levees, land cover and vegetation types, public recreation uses, and whether the property is in an incising or aggrading reach of the river.
Depth and extent of inundation at different river stages, frequency of inundation, and duration of inundation were significantly higher in informational value than flow velocities at different river stages (F 9,189 = 4.138, p < 0.001 Tukey's HSD p < 0.05).Depth and extent of inundation at different river stages, frequency of inundation, duration of inundation, and elevation were higher ranked in priority than flow velocities at different river stages and river stages associated with low flow events (F 9,175 = 9.691, p < 0.001; Tukey's HSD p < 0.05; Fig. 5A).There were similar differences in prioritization of inundation metrics when making management decisions in consideration of nonstationary conditions (e.g., climate change, F 9,190 = 9.79, p < 0.001; Fig. 5B).
The scientific products most useful to managers included layers in a geographical information system that would allow for dynamic interaction (100% of respondents), hydrologic model output files (65%), portable document format (PDF) files of maps for static interaction (55%), online map-viewer (35%), and smartphone application (25%).
Eighty percent of survey participants responded that their agency does not incorporate climate change into management plans or decisions for their floodplain lands.However, 70% of survey participants responded that they may have research needs pertaining to how climate change might influence decisions or objectives on the properties they manage.Of particular concern are changes to flow regimes, water level management, vegetation community changes, sedimentation, and groundwater changes.
When asked if managers had observed or have been informed of changes on their lands related to climate change, 25% of survey participants noted changes such as an increased occurrence of extreme precipitation events, increased base flows, a change in the timing of peak flows, expanding species ranges, or a change in timing of waterfowl migration.Ecology and Society 21(3): 12 http://www.ecologyandsociety.org/vol21/iss3/art12/DISCUSSION Floodplain conservation lands are managed for a variety of complex and competing purposes, yet our study found that the majority of resource managers agree that an understanding of inundation patterns is fundamental to manage multiple objectives and targets.In particular, depth and extent of inundation at different river stages, frequency of inundation, and duration of inundation were perceived as key metrics to understand across the floodplain for improving management outcomes.This understanding can be applied to determine disturbance regimes and habitat suitability for vegetation communities (Auble andScott 1998, Shafroth et al. 2002), moist soil units for waterfowl foraging habitat (Twedt 2013), and implications for fish nursery habitat (Sommer et al. 2001).Even with an understanding of the potential range of variability in river hydrogeomorphic conditions under climate change projections, managers maintain that depth, extent, frequency, and duration of inundation are the top priorities to understand for long-term management of floodplain conservation lands.
An increasing awareness of the role floodplain connectivity has on ecosystem functions and services has spurred interest in restoration of floodplain connectivity (Opperman et al. 2009, Schindler et al. 2014).Floodplain connectivity in this context includes surface-water and groundwater processes that result in exchange of organisms, nutrients, and organic matter between the main channel and the floodplain.Additionally, increased frequency of flooding in areas has raised concerns regarding a need to better understand flood risk (Gilles et al. 2012).Although flood-risk reduction was not a high priority objective in management of conservation lands, flood-risk reduction through flood water storage is an important ecosystem service that may benefit private land owners and urban centers located within the floodplain (Jacobson et al. 2015, Schober et al. 2015).As such, flood-risk reduction may be an underutilized argument in support of conservation management.
Understanding the relationships between inundation and ecological processes is an active and growing field of research because patterns of floodplain inundation are known to be primary drivers of ecological processes in floodplain ecosystems.For example, the ability to map areas of inundation at regional scales has recently provided information on extent of inundation for different flow return intervals (Chojnacki et al. 2012, Theiling andBurant 2013).In ungaged areas, remotely sensed data are being increasingly used to understand temporal inundation dynamics (Ward et al. 2014, Thomas et al. 2015).Spatially-explicit inundation information can be linked to known species responses to identify suitable habitat for species of interest (Jacobson et al. 2011) and evaluate changes in habitat availability under different management and climate change scenarios (Matella and Merenlender 2015).Two-dimensional hydrodynamic models have been employed to identify appropriate species to plant in restoration planning (Leyer et al. 2012).Our surveys provide evidence that managers recognize the linkage between inundation and ecological processes.The emphasis on inundation metrics supports the need for conservation agencies to invest in additional expertise in hydrology and geomorphology, or alternatively, to develop access to the expertise through collaborations with other agencies or institutions (Vaughan et al. 2009).
Understanding the influence of nonstationarity, in the form of climate change, on inundation metrics has significant implications for the management of floodplain lands.Although management plans for floodplain conservation lands rarely included more than a mention of climate change, shifting climate trends have been noted and acknowledged by many natural resource managers.The majority of resource managers surveyed have interest in understanding how climate is likely to change over time, and these changes may have dramatic effects to the biota in these riverfloodplain systems (Paukert and Galat 2010).As one survey participant responded, "If climate change is going to affect the properties I manage before I retire, I want to know about it."Recent U.S. Presidential executive orders (The White House 2013) guide and encourage all federal agencies to address climate change adaptation, however a lack of management-relevant climate change science, time, and funding remain obstacles to implementing adaptation (Kemp et al. 2015).
Our survey also revealed that time, funding, and a perceived disconnect between research and management communities limited the ability of resource managers to use new scientific information in management decisions.Kemp et al. (2015) found that federal agencies tasked with incorporating climate chance science into management of public lands were faced with similar limitations.Similarly, lack of information at relevant scales and budget constraints were the most limiting constraints for federal public land managers to adapt to climate change in mountainous Colorado (Archie et al. 2014).Although resource limitations of time and funding are often controlled by forces outside the realm of influence of individual managers and scientists, disconnections between research and management communities is a constraint that both resource managers and scientists can work to remove through regular communication and collaboration.The U.S. Forest Service has attempted to overcome this obstacle through the creation of climate change coordinator positions (Kemp et al. 2015) and science-management partnerships (Littell et al. 2012).The participating managers in our workshop were eager to help scientists understand the difficulties inherent to managing floodplains, and to aid in the identification of science products that meet their needs, suggesting partnerships between managers and researchers can identify science needs, specify relevant temporal and spatial scales, and determine user-friendly product formats.
Our survey results also reveal additional information gaps that may be limiting managers' ability to effectively manage for objectives such as controlling invasive species, maintaining respectful relationships with neighboring land owners, managing native, nongame species, managing endangered and threatened species, and promoting nutrient cycling.Similarly, there was limited information relative to priority of several conservation targets including pollinators, marsh birds, reptiles, shore birds, aquatic invertebrates, and amphibians.These limitations suggest that monitoring and pilot studies may be needed to gain information on floodplain ecosystem functions, habitat needs, and the role of management actions on these objectives and targets.Although many of the objectives and targets identified as information-limiting were also identified as low priorities, the priority ranking may be, at least partially, a result of limited understanding of how to effectively manage for an objective or target.In the case of pollinators, limited information available to http://www.ecologyandsociety.org/vol21/iss3/art12/understand recent population crashes (Vanbergen and Insect Pollinators Initiative 2013) has heightened interest and investment in pollinator research (Pollinator Health Task Force 2015).Along the same lines, a Midwest Marsh Bird Working Group was formed in 2012 to support regional marsh bird conservation through understanding impacts of management actions on marsh bird populations (Larkin et al. 2013).
Maintaining positive relationships with neighboring landowners was the highest scored priority across the study area.Many of the private lands in large river floodplains are in agricultural production and, therefore, landowners often have a different set of needs from the river than floodplain conservation land managers.Although the focus of our survey was on public lands, private lands make up a majority of floodplains in the Midwest as well as other large rivers in the world.The prevalence of privately-held agricultural lands in the floodplain and how those lands are managed is reflected in a 2012 survey of large river scientists and stakeholders who overwhelming (94%) agreed that current floodplain management does not support river-ecosystem needs (K.Lubinski and S. Gillespie, U.S. Geological Survey, unpublished data).Several of our survey participants noted the importance of determining the needs and values of private landowners and using that information to better link public and private land conservation programs.Research on private landowner management practices and values are limited, however, a survey of private landowners within the floodplain of the Missouri River in central Missouri identified rowcrop production, leaving land for children and grandchildren, soil conservation, and scenic beauty as the most important short-and long-term goals (Treiman and Dwyer 2004).Understanding the values of private land owners in the floodplain, as well as the larger community of people who live near rivers and their floodplains, could help natural resource managers communicate the optimal management actions to implement on public lands to support those values and also aid in developing novel approaches to encourage conservation practices on private lands (Raymond et al. 2015).
Given the complexities of managing floodplain conservation lands resulting from river-floodplain interactions, anthropogenic modifications, and the mosaic of private lands in the landscape, our research highlights the need for a few, relatively simple metrics of inundation to better manage these lands.The inundation metrics identified are evidence for the need to understand key drivers, i.e., hydrology, of river-floodplain ecosystems and, therefore, are likely related to the various objectives and conservation targets for which floodplain conservation lands are managed.

Responses to this article can be read online at
Maintain respectful relationships with neighbors Q17 Based on your experience, please rate the level of scientific information available on each objective (in general, not just on the identified properties).
Q20 Based on your experience, please rate the level of scientific information available on each of the additional objectives (in general, not just on the identified properties).

No opinion
Severely limited information available

Adequate information available
Objective 1 Q21 Please indicate the priority and management intensity of the following conservation targets for the property or majority of the properties you selected.
Q22 Based on your experience, please rate the level of scientific information available on each objective (in general, not just for the properties selected).

No opinion
Severely limited information available

Adequate information available
Pallid sturgeon Q23 Are there other conservation targets for the selected properties that we missed? Yes (Go to Q24)  No (Go to Q26) Q24 Please identify any additional conservation targets and indicate their priority and management intensity of the following conservation targets for the property or majority of properties you selected.
Q25 For any additional conservation targets, please rate the level of scientific information available on each objective, based on your experience (in general, not just for the properties you selected).
Q32 Please rate the following constraints based on how they limit your ability to obtain or utilize new scientific information in management decisions.Q7 What type of product output(s) from this project would you find useful?
 Environmental variables or indices (i.e., water, depth, and soil type) as layers in a Geographical Information system (i.e., ArcGIS) that would allow for dynamic interaction  Environmental variables or indices (i.e., water, depth, and soil type) as PDF/paper maps that would allow for static interaction  Biological model outputs

Fig. 1 .
Fig. 1.Floodplain conservation lands within the Upper Mississippi River (UMR), Middle Mississippi River (MMR), and Lower Missouri River (LMOR) were the focus of this research.

Fig. 3 .
Fig. 3. Average survey responses (N = 54) across the study area for management priority and level of scientific information available to manage for selected (A) objectives and (B) conservation targets on floodplain conservation lands.Dashed lines are breakpoints to differentiate between high and low priorities and sufficient and limited amounts of information available.

Fig. 4 .
Fig. 4. Information available relative to priority for (A) management objectives and (B) conservation priority.Values on x-axis were calculated by first standardizing priority and information available scores, then subtracting the standardized priority value from the standardized information available value.

Fig. 5 .
Fig. 5. Boxplots of rankings from 21 respondents to the followup survey on the development priority of various metrics identified as science needs in our workshop for (A) current management and (B) management under projected climate changes.An analysis of variance and Tukey's HSD test was performed on each dataset to determine statistical differences in mean rank between metrics.Letters within each box that differ indicate ranks that differ at alpha = 0.05.Metrics (see Appendix 3 for full description of each metric) included depth and extent of floodplain inundation at different river stages, frequency of inundation, duration of inundation, elevation, seasonality of inundation, groundwater table associated with low and high flows, extent of seep-driven and precipitationdriven ponding on "dry" side of levee, locations of high probability of sedimentation or scour, stages associated with low flow events, and flow velocities on floodplain at different stages.
: http://www.ecologyandsociety.org/issues/responses.php/8620 Q1 Please select the management agency you currently work for:  US Army (Go to Q2)  US Fish and Wildlife Service (Go to Q3)  US Forest Service (Go to Q4)  US National Park Service (Go to Q5)  Illinois Department of Natural Resources (Go to Q6)  Iowa Department of Natural Resources (Go to Q7)  Kansas Department of Wildlife, Parks, and Tourism (Go to Q8)  Minnesota Department of Natural Resources (Go to Q9)  Missouri Department of Conservation (Go to Q10)  Missouri Department of Natural Resources (Go to Q11)  Nebraska Game and Parks Commission (Go to Q12)  Wisconsin Department of Natural Resources (Go to Q13)  Other (Go to Q14) ____________________ Q2 Please identify the U.S. Army properties of which you are actively engaged in managing (conservation lands within the floodplain of the Lower Missouri River and Upper and Middle Mississippi River only


Hydrologic model outputs  Online mapviewer  Smartphone or tablet application  Other ____________________ Q8 Does your agency incorporate climate change into management plans for floodplain properties? Yes  No Q9 Does your agency incorporate climate change into management decisions on floodplain properties? Yes  No Q10 Do you have research needs pertaining to how climate change might influence management decisions or objectives on the floodplain properties you manage? Yes  No  Not sure Q11 Please describe your research needs pertaining to how climate change might influence management decisions or objectives on the properties you manage.Q12 Have you seen or been informed of changes or trends associated with climate change on the properties you oversee? Yes  No Q13 Please identify the changes or trends on these properties that you associate with climate change.Q14 In light of the information available regarding trends and projections in climate, hydrology, and extreme floods and droughts, please rank the priority in which you would like to see the following metrics developed (1 = highest priority).Add additional variables you would find useful to make management decisions regarding climate change impacts on floodplain properties.provide additional feedback that you would like to share regarding the objectives of the survey or specific questions.

Table 1 .
Defining attributes of the river-floodplain dynamics of the three study reaches.

Table 2 .
Lists of selected management objectives, conservation targets, information sources, and constraints in floodplain management included in the first survey to 80 floodplain managers in the Lower Missouri and Upper and Middle Mississippi rivers.
management, of different types of scientific information in decision making.The follow-up survey also included questions on how management agencies incorporate climate change, a source of nonstationarity, into management plans and what types of scientific products and formats would be most useful for transferring knowledge to managers.Both online surveys were developed and distributed using Qualtrics (http://www.qualtrics.com/) and reviewed and approved by the University of Missouri Institutional Review Board (Project #1213966).
). Please identify the U.S. Forest Service properties of which you are actively engaged in managing (conservation lands within the floodplain of the Lower Missouri River and Upper and Middle Mississippi River only).Please identify the Illinois Department of Natural Resources properties of which you are actively engaged in managing (conservation lands within the floodplain of the Upper and Middle Mississippi River only). Big River State Forest  Cape Bend State Fish and Wildlife Area  Cedar Glen State Natural Area  Chouteau Island State Fish and Wildlife Area  Delabar State Park  Devil's Island State Fish and Wildlife Area  Elton Hawks Bald Eagle Refuge Nature Preserve  Henderson Creek State Fish and Wildlife Area  Horseshoe Lake Nature Preserve  Mississippi River Sand Hills Natural Area  Mississippi River State Fish and Wildlife Area  Union County State Fish and Wildlife Area  Other ____________________ Go to Q16 Q7 Please identify the Iowa Department of Natural Resources properties of which you are actively engaged in managing (conservation lands within the floodplain of the Lower Missouri River and Upper and Middle Mississippi River only).Please identify the Kansas Department of Wildlife, Parks, and Tourism properties of which you are actively engaged in managing (conservation lands within the floodplain of the Lower Missouri River only).Please identify the Missouri Department of Natural Resources properties of which you are actively engaged in managing (conservation lands within the floodplain of the Lower Missouri River and Upper and Middle Mississippi River only).Please identify the Wisconsin Department of Natural Resources properties of which you are actively engaged in managing (conservation lands within the floodplain of the Upper Mississippi River only).
Q33 Are there other constraints that limit your ability to obtain or utilize new scientific information in management decisions? Yes (Go to Q34)  No (Go to Q35) Q34 Please identify additional constraints and rate them based on how they limit your ability to obtain or utilize new scientific information in management decisions.Please share the source of environmental data, type of data collected, and location of collection sites.Q6 Please rank the potential informative value of the following types of scientific information in relation to managing floodplain properties.Additionally, please rank the priority in which you would like to see the following metrics developed (1 = highest priority).Add additional variables you would find useful to make management decisions on floodplain properties.