Reducing nitrogen delivery to coastal waters is a 'wicked problem' that involves trade-offs among environmental, economic and social need domains. Because these trade-offs arise from spatial and temporal complexities in sources and sinks of this element, we hypothesized that a transdisciplinary focus on disproportionality could enable the identification of 'sweet spots' where multiple factors converge to create opportunities to control nitrogen flux. We applied this approach to the region of Baltimore, MD, USA, by mapping stream reaches with high nitrogen concentrations, hydrologic conditions that facilitate high rates of nitrogen removal following stream restoration, a high household willingness to pay for restoration projects and a high social need for restoration, and subsequently identifying locations where these factors converge to create sweet spots. Our analysis indicates that sweet spots that optimize environmental, economic and social need components of sustainability may be rare in cities across the USA. The desire to bundle multiple benefits in the budgeting for environmental interventions such as stream restoration may create a suboptimal distribution of these interventions.

Water resources planning is dependent on hydrologic models to estimate flows and storage in candidate engineering designs. However, such models are calibrated with limited flow data relative to the many model parameters. This may result in different equifinal parameterizations that imply different optimal designs. To assess if and how this uncertainty should be considered, we compare three methods for multiobjective optimization of green infrastructure (GI): one that designs to the most likely parameterization and two robust alternatives that use several likely parameterizations with (1) likelihood-weighted objective functions, and (2) min-max objective functions. To evaluate these methods, we set synthetic true values for model parameters, use them to simulate observed streamflow, and then use Bayesian calibration to estimate parametric uncertainty. We compare results from optimization to the synthetic parameterization against the three alternatives. The GI optimizations aim to minimize flooding, low-flow intensification, and cost. We find that the two robust methods provide objective values and decisions that are closer to those optimized to the synthetic truth, demonstrating value in considering hydrologic model uncertainty in water resource system designs.

Forests are crucial for the production of high-quality freshwater resources. Complex interactions between climate change and forest processes can result in uncertainty in the availability of freshwater to downstream communities and the environment. Previous studies reported consistent increasing trends in global river discharge during the last century, which has been explained by either climate factors (usually called “hydrological intensification”) or suppressed transpiration due to CO2-induced stomatal closure. In this study, we study long-term changes in hydrological partitioning of precipitation between evapotranspiration and runoff generation (as mm per year) along a gradient of forested watershed along the eastern temperate forest biome. The precipitation is increasing at faster rates than runoff at most of these study catchments, which suggests long-term increases in evapotranspiration. These divergent trends in precipitation versus runoff rates are significantly correlated to long-term trends in NDVI and growing season length at the watershed scale, while climate variables cannot provided significant explanation. These findings suggest that the combined effect of increased temperatures and CO2 fertilization have led to increased leaf area and lengthened growing season, which may counteract the effect of the CO2-induced stomatal closure across the eastern US. This study emphasizes the importance of understanding vegetation responses to climate change to predict future flow regimes in forested watersheds.

Reproducible environmental modelling often relies on spatial datasets as inputs, typically manually subset for specific areas. Yet, models can benefit from a data distribution approach facilitated by online repositories, and automating processes to foster reproducibility. This study introduces a method leveraging diverse state-scale spatial datasets to create cohesive packages for GIS-based environmental modelling. These datasets were generated and shared via GeoServer and THREDDS Data Server connected to HydroShare, contrasting with conventional distribution methods. Using the Regional Hydro-Ecologic Simulation System (RHESSys) across three U.S. catchment-scale watersheds, we demonstrate minimal errors in spatial inputs and model streamflow outputs compared to traditional approaches. This spatial data-sharing method facilitates consistent model creation, fostering reproducibility. Its broader impact allows scientists to tailor the method to various use cases, such as exploring different scales beyond state-scale or applying it to other online repositories using existing data distribution systems, eliminating the need to develop their own.

Efforts to reduce nitrogen and carbon loading from developed watersheds typically target specific flows or sources, but across gradients in development intensity there is no consensus on the contribution of different flows to total loading or sources of nitrogen export. This information is vital to optimize management strategies leveraging source reductions, stormwater controls, and restorations. We investigate how solute loading and sources vary across flows and land-use using high frequency monitoring and stable nitrate isotope analysis from five catchments with different sanitary infrastructure, along a gradient in development intensity. High frequency monitoring allowed estimation of annual loading and attribution to storm versus baseflows. Nitrate loads were 16 kg/km2/yr. from the forested catchment and ranged from 68 to 119 kg/km2/yr., across developed catchments, highest for the septic served site. Across developed catchments, baseflow contributions ranged from 40% of N loading to 75% from the septic served catchment, and the contribution from high stormflows increased with development intensity. Stormflows mobilized and mixed many surface and subsurface nitrate sources while baseflow nitrate was dominated by fewer sources which varied by catchment (soil, wastewater, or fertilizer). To help inform future sampling designs, we demonstrate that grab sampling and targeted storm sampling would likely fail to accurately predict annual loadings within the study period. The dominant baseflow loads and subsurface stormflows are not treated by surface water management practices primarily targeted to surface stormflows. Using a balance of green and gray infrastructure and stream/riparian restoration may target specific flow paths and improve management. Baseflows contributed nearly as much or more of nitrogen export as stormflows across low to moderate development intensity catchments Stormflow NO3- sources were diverse and highly variable in time, while baseflows mobilized few sources with little temporal variability Grab sampling methods did not reliably estimate the loading observed from in situ monitoring for small catchments

Excess export of reactive nitrogen in the form of nitrate (NO3-) from suburban watersheds is a major source of water quality degradation and threatens the health of downstream and coastal waterbodies. Ecosystem restoration and best management practices (BMPs) can be introduced to reduce in-stream NO3- loads by promoting vegetation uptake and denitrification in the upland and riparian areas. However, accurately evaluating the effectiveness of these practices and setting regulations for nitrogen inputs requires an understanding of how human sources of nitrogen interact with ecohydrological systems. We evaluated how the spatial and temporal distribution of nitrogen sources interacts with ecohydrological transport and transformation processes along surface and subsurface flow paths with respect to nitrogen cycling and export. Embedding distributed household sources of nitrogen and water within hillslope hydrologic systems influences the development of both planned and unplanned "hot spots" of nitrogen flux and retention in suburban ecosystems. We chose a well-monitored low-density suburban watershed, Baisman Run, in Baltimore County, Maryland, USA, to evaluate patterns of in-stream NO3- concentrations and terrestrial nitrogen cycling processes in response to three common activities: irrigation, fertilization, and on-site sanitary wastewater disposal (septic systems). We augmented a distributed ecohydrological model, RHESSys (Regional Hydro-Ecological Simulator System), with estimates of the spatial distribution of these loads at household parcel level to develop a predictive understanding of the factors generating upland and riparian nitrogen cycling, transport, and stream NO3- concentrations. We calibrate subsurface hydraulic parameters only without calibrating ecosystem and biogeochemical processes. The calibrated model predicted mean NO3- concentrations of 1.43 mg NO3--N L-1 compared to the observed 1.6 mg NO3--N L-1 from water year 2013 to 2017. With spatially explicit irrigation, fertilizer, and septic effluent inputs, estimated denitrification rates in grass lawns, a dominant land cover in suburban landscapes, were also in the range of previously measured values. The highest predicted denitrification rates (N retention hot spots) were downslope of lawn and septic locations in a constructed wetland and at a riparian sediment accumulation zone at the base of a gully receiving street drainage. These locations illustrate the development of hot spots for nitrogen cycling and export in both planned and "accidental" retention features. Appropriate siting of suburban nutrient management and BMPs should assess and incorporate spontaneously developed nutrient hot spots to design improved landscape ecosystem N retention and water quality.

Forests play a key role in the mitigation of global warming and provide many other vital ecosystem goods and services. However, as forest continues to vanish at an alarming rate from the surface of the planet, the world desperately needs knowledge on what contributes to forest preservation and restoration. Migration, a hallmark of globalization, is widely recognized as a main driver of forest recovery and poverty alleviation. Here, we show that remittance from migrants reinforces forest recovery that would otherwise be unlikely with mere migration, realizing the additionality of payments for ecosystem services for China's largest reforestation policy, the Conversion of Cropland to Forest Program (CCFP). Guided by the framework that integrates telecoupling and coupled natural and human systems, we investigate forest-livelihood dynamics under the CCFP through the lens of rural out-migration and remittance using both satellite remote sensing imagery and household survey data in two representative sites of rural China. Results show that payments from the CCFP significantly increases the probability of sending remittance by out-migrants to their origin households. We observe substantial forest regeneration and greening surrounding households receiving remittance but forest decline and browning in proximity to households with migrants but not receiving remittance, as measured by forest coverage and the Enhanced Vegetation Index derived from space-borne remotely sensed data. The primary mechanism is that remittance reduces the reliance of households on natural capital from forests, particularly fuelwood, allowing forests near the households to recover. The shares of the estimated ecological and economic additionality induced by remittance are 2.0% (1.4%∼3.8%) and 9.7% (5.0%∼15.2%), respectively, to the baseline of the reforested areas enrolled in CCFP and the payments received by the participating households. Remittance-facilitated forest regeneration amounts to 12.7% (6.0%∼18.0%) of the total new forest gained during the 2003-2013 in China. Our results demonstrate that remittance constitutes a telecoupling mechanism between rural areas and cities over long distances, influencing the local social-ecological gains that the forest policy intended to stimulate. Thus, supporting remittance-sending migrants in cities can be an effective global warming mitigation strategy.

Hydrologic alterations associated with urbanization can weaken connections between riparian zones, streams, and uplands, leading to negative effects on the ability of riparian zones to intercept pollutants carried by surface water runoff and groundwater flow such as nitrate (NO3 -) and phosphate (PO4 3-). We analyzed the monthly water table as an indicator of riparian connectivity, along with groundwater NO3 - and PO4 3- concentrations, at four riparian sites within and near the Gwynns Falls Watershed in Baltimore, MD, from 1998 to 2018. The sites included one forested reference site (Oregon Ridge), two suburban riparian sites (Glyndon and Gwynnbrook), and one urban riparian site (Cahill) with at least two locations and four monitoring wells, located 5 m from the center of the stream, at each site. Results show an increase in connectivity as indicated by shallower water tables at two of the four sites studied: Glyndon and Cahill. This change in connectivity was associated with decreases in NO3 - at Glyndon and increases in PO4 3- at Glyndon, Gwynnbrook, and Cahill. These changes are consistent with previous studies showing that shallower water table depths increase anaerobic conditions, which increase NO3 - consumption by denitrification and decrease PO4 3- retention. The absence of change in the forested reference site, where climate would be expected to be the key driver, suggests that other drivers, including best management practices and stream restoration projects, could be affecting riparian water tables at the two suburban sites and the one urban site. Further research into the mechanisms behind these changes and site-specific dynamics is needed.

Identifying locations prone to exporting nitrogen (N), also called export control points, within residential landscapes, is key to determining N mitigation strategies. Within residential landscapes, lawns have the potential to act as either a sink of N via uptake and denitrification, or a source of N via additions such as fertilizer. Lawns draining to impervious surfaces are more likely to be sources of N loading to receiving water bodies through directly connected curb and sewer flow paths. We utilized small-scale rainfall experiments to examine whether hydrobiogeochemical measurements of potential denitrification and saturated infiltration rates were predictive of N mobilization, and how potential export control points (locations within the upper quartile of N mobilization values) varied spatially and temporally on residential lawns in Baltimore, Maryland. We found potential denitrification, but not infiltration, was predictive of N mobilization in runoff and leachate, only on fertilized lawns. Potential export control points occurred more often in the late summer and fall and 85% were on fertilized lawns. Applying fertilizer shortly before a rainfall event increased the N mobilization in runoff and leachate by an order of magnitude. Suburban front yards also had more potential export control points compared to backyards, which is notable as front yards are surrounded by impervious surfaces increasing their vulnerability to transporting N to downstream ecosystems. These findings highlight the spatial and temporal variability of N mobilization on lawns. Targeting locations such as vulnerable front yards, or behaviors, such as timing of fertilizer application, may be useful N mitigation strategies.

Urban watersheds have experienced ecosystem degradation due to land cover change from vegetation to impervious areas. This transformation results in increased stormwater runoff, stream channel erosion and sedimentation, and both increased inputs and reduced ecosystem retention of nutrients. Ecosystem restoration practices, including terrestrial and aquatic low impact development (LID), are becoming widely implemented in urban watersheds globally. A major question is how "green" and "grey" infrastructure can be optimally balanced to shift ecohydrological behavior towards pre-urbanization conditions. Traditional stormwater engineering typically controls runoff by temporary storage (detention) and release of stormwater, while LID designs are developed to reduce runoff by a combination of infiltrating precipitation and evapotranspiration, while promoting biogeochemical retention of nutrients. These practices are often combined with stream and riparian restoration that increases nutrient retention and reduces in-stream loads. In this study, we simulated the potential impact of three types of terrestrial LID and green infrastructure (GI) on watershed runoff and nitrate (NO3-) loading to local streams, independent of detention storage effects. The treatments included increased tree canopy, vegetated roadside bioswales, and permeable pavement. We then evaluated the individual and interactive impacts of these practices on the effectiveness of NO3- load reduction provided by stream restoration, which is affected by the altered runoff and nutrient loading caused by the LID and GI. Urban reforestation provided the highest effectiveness in terms of reducing stormflow and nutrient export, while bioswales and permeable pavement unexpectedly increased in-stream NO3- loads. Retrofit of the previously developed watershed by LID/GI alone may not provide sufficient mitigation in stormwater and nutrient loads, and should be balanced with additional grey infrastructure, such as detention ponds, rain cisterns, and sewer system upgrades.

Recent studies have demonstrated some advantages of using advanced heuristic algorithms to identify near-Pareto-optimal future locations, types, and sizes for stormwater low-impact development and green infrastructure (LID/GI) across a given urban landscape. However, previous optimization studies did not consider social equity as an objective, which poses problems because urban green infrastructure often is distributed inequitably. Increasing access to LID/GI in historically marginalized areas is a prominent environmental justice issue, and increasingly is becoming a primary consideration when prioritizing future locations, types, and sizes of urban LID/GI. This study integrated a novel spatial social equity objective [LID/GI-Social Vulnerability Index (SVI) correlation objective, ?] into a multiobjective LID/GI optimization model. The LID/GI-SVI correlation is an objective that directs the optimization algorithm to search for LID/GI distributions that maximize the linear correlation between LID/GI implementation and subbasins with higher estimated percentages of historically marginalized people. Our analysis focused on understanding the impacts of the LID/GI-SVI correlation objective on a LID/GI optimization model. This modeling study demonstrates that (1) the LID/GI-SVI correlation objective can be used to direct optimization algorithms to search for LID/GI distributions that can achieve runoff management objectives, increase green LID/GI implementation in more marginalized areas, and explore the potential trade-offs or synergies between hydrologic and equity goals; (2) LID/GI optimization formulations that consider only hydrologic objectives likely will not result in equitable LID/GI distributions; (3) LID/GI distributions that perform well on the LID/GI-SVI correlation may be composed of different types of LID/GI than less-equitable but more hydrologically favorable LID/GI distributions; and (4) for our study area, including spatial equity as an objective resulted in modest reductions in the hydrologic performance of near-Pareto-optimal LID/GI distributions.

Nonpoint source (NPS) water quality trading (WQT) is a market-based approach to improving water quality. Past work has shown that these programs could increase localized pollutant loadings, in part by exporting water quality controls from urban to rural areas. Virginia's NPS WQT program has enabled thousands of transactions and may provide a model for other programs, but its impacts on urban water quality have not been thoroughly assessed. We quantify the impact of NPS WQT purchases in Virginia on water quality and hydrology in an urban catchment. We go on to assess outcomes of a policy alternative where buyers and sellers are collocated in the urban catchment. Simulation results show that NPS WQT increased total phosphorus (TP) loading by an average of 0.8 lbs TP/year for each 1.0 offsite credits purchased in the analyzed catchment. The TP loading increased in years with greater rainfall, such that TP loads were increased by up to 1.2 lbs TP/year for each offsite credit purchased. These loading increases may or may not be acceptable, depending on the cumulative number of purchases within an urban catchment and existing local water quality issues. In our policy alternative with buyers and sellers collocated in the catchment, we found that the TP increase from development was completely offset at the catchment scale, with a decrease of 4.3 lbs TP/year for each 1.0 credits purchased. This suggests that credits awarded for urban mitigation practices are undervalued compared with water quality requirements for credit purchasers. This undervaluation is a result of the Virginia trading program using one approach to compute the credit value for buyers and a different approach to compute the credit value for sellers. We demonstrate how using a single model to determine both buyer and seller credit values in urban areas could provide greater transparency and mitigate the risk of urban pollution hot spots. This work demonstrates the importance of consistency in the scale of pollutant load calculations between buyers and sellers for NPS WQT, and contributes novel insight into the implications of WQT for urban NPS pollution.

Much attention has been given to the disproportionate streamflow deficits (relative to rainfall deficits) experienced by many catchments during the Millennium Drought (1998–2009) in southeastern Australia, along with lack of post‐drought streamflow recovery in some cases. However, mechanisms behind the coupled hydrologic and ecosystem dynamics are poorly understood. We applied a process‐based ecohydrologic model (RHESSys) in a Melbourne water supply catchment to examine changes in ecohydrologic behavior during and after the drought. Our simulations suggested that average transpiration (green water) was maintained under drought despite a substantial (12%) decrease in average rainfall, meaning that the entire rainfall deficit translated to reduced streamflow (blue water). Altered spatial patterns of vegetation behavior across the terrain helped the ecosystem maintain this unexpectedly high green water use. Decreased transpiration upland was compensated by increases in the riparian zone, which was less water limited and therefore able to meet higher water demand during drought. In the post‐drought period, we found greater transpiration and reduced subsurface water storage relative to pre‐drought, suggesting a longer‐term persistence in altered water partitioning. The post‐drought outcome was attributed to a combination of warmer climate and the persisting effects of the drought on nutrient availability. Given the importance of shifting ecohydrologic patterns across space, our results raise concerns for applying lumped conceptual hydrologic models under nonstationary or extreme conditions. Additionally, the processes we identified have important implications for water supply in Australia's second largest city under projected drying.

Residential landscapes are essential to the sustainability of large areas of the United States. However, spatial and temporal variation across multiple domains complicates developing policies to balance these systems' environmental, economic, and equity dimensions. We conducted multidisciplinary studies in the Baltimore, MD, USA, metropolitan area to identify locations (hotspots) or times (hot moments) with a disproportionate influence on nitrogen export, a widespread environmental concern. Results showed high variation in the inherent vulnerability/sensitivity of individual parcels to cause environmental damage and in the knowledge and practices of individual managers. To the extent that hotspots are the result of management choices by homeowners, there are straightforward approaches to improve outcomes, e.g. fertilizer restrictions and incentives to reduce fertilizer use. If, however, hotspots arise from the configuration and inherent characteristics of parcels and neighborhoods, efforts to improve outcomes may involve more intensive and complex interventions, such as conversion to alternative ecosystem types.

The lines between natural areas and human habitats have blurred as urbanization continues, creating a need for the study of ecosystems at all levels of development. This need is particularly acute for exurban environments, which have low population density but are rapidly changing and have a dynamic mix of natural and human-dominated features. We examined long-term (1998–2018) trends in nitrate and chloride concentrations and fluxes in forested and exurban streams in Baltimore County, Md USA. Concentrations and fluxes of nitrate and chloride were an order of magnitude higher in the exurban stream than the forested stream and were increasing even though snowfall and road salt use did not increase over the study period. In the forested stream, concentrations and fluxes of chloride increased from 1998 to 2008, but decreased from 2008 to 2018 due to unquantified factors. Concentrations of nitrate decreased in the forested stream, likely due to decreases in atmospheric deposition. These decreases in atmospheric deposition, and efforts to reduce fertilizer use by county governments, do not appear to have affected nitrate concentrations and fluxes in the exurban watershed. Any efforts to reduce the concentrations and fluxes of chloride and nitrate in exurban streams will likely benefit substantially from further understanding of the mechanisms underlying the temporal patterns.

In this study, we investigate how seasonal streamflow and soil moisture patterns have responded to variability in vegetation phenology in humid, temperate forested watersheds without significant seasonal snowmelt over the last four decades. We characterize spring streamflow peaks using 50th percentiles of cumulative daily precipitation, streamflow, and soil moisture measurements, and investigate interactions with remotely sensed, greenup anomalies. After removing a dominant precipitation control, 1-day earlier greenup is usually associated with about 1-day early spring flow peak at four low-elevation deciduous catchments using both sequential and multiple linear regressions. This indicates that the strong dependency of seasonal flow regimes on precipitation is mediated by vegetation seasonality, especially by greenup variability. In contrast, we find less significant correlations of the greenup anomalies on flow percentiles from two paired evergreen and two high-elevation deciduous catchments. At a plot scale, similar correlations were found only at an upslope topographic position, where precipitation also showed tighter coupling with moisture seasonal patterns than downslope. Our study suggests that rainfall-runoff and rainfall-soil moisture relations have been closely mediated by vegetation seasonality in deciduous forests, especially by greenup anomalies, but patterned along topoclimate and hillslope gradients. This study emphasizes that it is important to understand phenological responses to ongoing climate change (in both long-term and interannual variability) for prediction of seasonal flow regimes especially in deciduous forested catchments. Topoclimate modulates coupling of precipitation and greenup with seasonal runoff dynamicsDeciduous catchments show tighter coupling between greenup and seasonal runoff dynamics than paired evergreen onesHillslope modulates coupling of precipitation and greenup with seasonal soil moisture dynamics

Ongoing warming under climate change generally brought lengthened growing season and subsequent changes in evapotranspiration (ET) seasonality, but its implication in future freshwater regimes is not fully understood. In this study, we investigate how remotely sensed vegetation phenology information has correlated with seasonal flow patterns in the southern Appalachians over the last three decades. We characterize low-frequency seasonal peaks using cumulative 50 percentiles of daily precipitation, streamflow and soil moisture measurements, further related to long-term greenup anomalies in the study site. After removing a dominant precipitation control, the greenup anomalies show a significant correlation with spring flow peak timing at a low-elevation deciduous catchment. This indicates strong dependency of seasonal flow regimes on precipitation was mediated by warming-induced extended growing season, especially by early greenup. At a plot scale, similar correlations were found only at an upslope topographic position, which occupies most of forest landscapes within the catchment. Meanwhile, we found less significant correlations of the greenup anomalies on the flow percentiles at an adjacent evergreen and a high-elevation deciduous catchments. This suggests that early greenup in deciduous forests leading to early ET increases, in turn result in early soil moisture dry-down patterned by hillslope positions, and early seasonal streamflow declines given precipitation. Our study suggests that it is likely that spring flow peak timing in this region continue to be shifted earlier with time by warming-induced early greenup, which would have great implications for understanding future freshwater availability in the southeastern US.

Nonpoint source urban nutrient loading into streams and receiving water bodies is widely recognized as a major environmental management challenge. A dominant research and management paradigm assumes that loading primarily derives from elevated stormwater. However, baseflow can account for a large portion of total loading, especially for low development intensity watersheds which comprise the largest urban areas. We investigated the sources and drivers of nonpoint source baseflow nitrogen loading across 27 headwater catchments in the urbanized Piedmont region of North Carolina, USA. Nitrate isotopes, predictors of concentration‐discharge (CQ) slopes, and predictors of mean annual total dissolved nitrogen (TDN) loading suggest that wastewater was a major baseflow nonpoint source of nitrogen across developed catchments likely contributing 61% of nitrate loading from septic served catchments and 49% from sewer served catchments. Our findings suggest that subsurface TDN was abundant, loading was largely transport limited, and the hydrogeomorphic position of sanitary infrastructure strongly influences transport. We developed an empirical model showing catchment loading increased with the topographic wetness index of sanitary sewer location, convergent sloping land area, parcel density, and residual agricultural landcover (R2 = 0.78). We extended this model to the study region's 1,436 developed small (0.3–20.8 sq km) catchments. We estimated up to 92.7% of nonpoint source baseflow TDN loading comes from low and medium development intensity catchments, and sanitary infrastructure in wet areas of the landscape accounts for 39% of regional baseflow loading. Our research indicates that managing baseflow loading will require addressing lower development intensity catchments and sanitary infrastructure.

It is common to test hydrologic models under contrasting historical periods as an indicator of likely performance under climate change. For example, a model calibrated under average conditions may be tested under increasingly dry subsets of the observational record. Any decline in performance as the testing conditions deviate further from the calibration conditions is then assumed to represent likely performance degradation under climate change scenarios with comparable rainfall decreases. Many studies have inherently applied the assumption that past rainfall variability can be used as a proxy for future climate change, but the analogy may be flawed for three main reasons: * Due to lagged hydrologic response to meteorological shifts, catchment behaviour under long-term wetting or drying may not be fully represented over shorter wet or dry periods. * Subsets of the past record selected based on rainfall are unlikely to reflect future temperature increases. * Past observations do not include expected increases in carbon dioxide levels. If any of these factors substantially impacts catchment response, subsets of the historical record with equivalent rainfall will not be accurate proxies for future climate scenarios. We tested the impact of each factor using the ecohydrologic model RHESSys. RHESSys dynamically simulates vegetation growth, subsurface flow and nutrient cycling and is thus able to capture the key processes that could drive nonstationary catchment response in the future. We found that all three future climate factors (rainfall change persistence, temperature, and carbon dioxide) altered catchment response substantially, especially for drier future scenarios. For our study catchment, persistence of dry conditions over many decades led to different subsurface water storage levels than the same rainfall experienced over shorter timeframes, leading to different streamflow. The impacts of increased temperature and carbon dioxide concentrations on vegetation further altered runoff behaviour. This means that long-term climate change effects will not necessarily emerge over short historical periods with equivalent rainfall. In our example, ignoring persistence in rainfall changes, rising temperatures, and higher carbon dioxide levels could lead us to underestimate model performance degradation in terms of Nash-Sutcliffe efficiency by as much as 0.41. Therefore, the uncertainty introduced in hydrologic models by future climate change has probably been underestimated in the current literature.

Stream restoration is widely used to mitigate the degradation of urban stream channels, protect infrastructure, and reduce sediment and nutrient loadings to receiving waterbodies. Stabilizing and revegetating riparian areas can also provide recreational opportunities and amenities, and improve quality of life for nearby residents. In this project, we developed indices of an environmental benefit (potential nitrate load reduction, a priority in the Chesapeake Bay watershed) and economic benefit (household willingness to pay, WTP) of stream restoration for all low order stream reaches in three main watersheds in the Baltimore metro region. We found spatial asynchrony of these benefits such that their spatial patterns were negatively correlated. Stream restoration in denser urban, less wealthy neighborhoods have high WTP, but low potential nitrate load reduction, while suburban and exurban, wealthy neighborhoods have the reverse trend. The spatial asynchrony raises challenges for decision makers to balance economic efficiency, social equity, and specific environmental goals of stream restoration programs.