Removing the two clustered cropland playas from the analyses did not influence statistical significance at α = 0.1, therefore results are presented for the model including all sample sites in the RWB.

There were no interactions among fixed effects for any GHG emissions in RWB playas. There was seasonal variation in all GHG emissions with the bulk of emissions occurring during peak growing season. Emissions of CO₂ differed among land use types but not by landscape position or by year. Cropland playas emitted 28% more CO₂ than WRP playas; reference playas did not differ from either cropland or WRP playas (Table 1). Methane fluxes differed among landscape positions but did not differ by land use type or by year. As expected, CH₄ emissions were highest in the wetland center and decreased at each landscape position moving outward to the wetland edge (Table 2). Emissions of N₂O did not differ among land use types, landscape positions or by year. Net emissions differed among land use types and landscape positions but not by year. Cropland playas emitted 53% more CO_2-equiv than WRP playas; reference playas did not differ from either cropland or WRP playas (Figure 1(a)). The bulk of emissions occurred at the center of the wetland which differed from both the wetland edge and 1/2 distance between center and edge, which did not differ from one another.

There were no interactions among fixed effects for soil moisture and soil temperature in RWB playas. Seasonal variation existed for soil moisture which

was generally highest early in the season, coinciding with snow melt and spring rains. Soil moisture also differed among land use types, landscape positions, and by year. Playas in WRP had 19% lower soil moisture than cropland and reference playas, which did not differ (Table 3). Soil moisture was also lower at the wetland edge than at either point in the wetland basin; wetland basin landscape positions (positions 1 and 2) did not differ from one another. Finally, there was an 8% increase in soil moisture in 2013 compared to 2012. There was seasonal variation in soil temperature which was obviously highest during summer. Soil temperature also differed among land use types and by year but not among landscape positions. Playas in the WRP were approximately 1.4˚C warmer than reference and cropland playas, which did not differ from one another (Table 3). Also, playa soils were approximately 1.6˚C warmer in 2012 than in 2013.

Plant species richness did not differ among land uses in playas in 2013, however aboveground biomass did. Cropland playas had 30% more aboveground biomass than reference condition playas and 63% more than playas in WRP; reference condition playas had 35% more aboveground biomass than playas in WRP (Table 3).

There was no land use * year interaction for any GHG emissions in RWB uplands. There was seasonal variation in all GHG emissions with the bulk of emissions occurring during peak growing season. Emissions of CO₂ and CH₄ did not differ among land use types or by year. Emissions of N₂O differed among land use type and by year. Cropland uplands emitted 648% more N₂O than WRP and reference uplands (Table 1). In addition, twice as much N₂O was emitted in 2013 than 2012. Overall, net CO_2-equiv emissions did not differ among land use types or by year (Figure 1(b)).

There was no land use * year interaction for soil moisture or soil temperature in RWB uplands. Soil moisture differed among land use types and by year. Cropland uplands were 34% more moist than WRP uplands; soil moisture in reference uplands did not differ from cropland or WRP uplands (Table 3). Similar to wetlands, RWB uplands were 14% more moist in 2013 than 2012. Soil temperature differed among land use types with WRP uplands averaging 1.1˚C warmer than cropland and reference uplands, which did not differ (Table 3). In 2012, soils averaged 1.6˚C warmer than in 2013.

In 2013, plant species richness differed among land uses in adjacent uplands. As expected, upland croplands contained half as many plant species as reference condition and WRP uplands, which did not differ from one another. Aboveground biomass was similar among land uses (Table 3).

There was a land use type * year interaction for playa CO₂ emissions. Emissions of CO₂ did not differ within land use types between years for cropland and grassland playas but in CRP playas it was 29% lower in 2013 than 2012 (Table 1). Emissions of CO₂ did not differ among landscape positions. Fluxes in CH₄ differed by land use type * landscape position and land use type * year interactions. Wetland center contributed most to CH₄ emissions but only differed from wetland edge, except in CRP playas, where CH₄ emissions were negligible throughout the wetland (Table 2). Wetland edge and ½ distance between center and edge did not differ. There were no differences among land use types in 2012, however, in 2013, cropland and grassland playas emitted 46 and 23 times more CH₄, respectively, than playas in CRP (Table 1). Only cropland playas differed between years, changing from a small sink for CH₄ in 2012 to the greatest contributor of CH₄ emissions among land uses in 2013. There were no interactions among fixed effects for playa N₂O emissions. Emissions of N₂O differed among land use types but not among landscape positions or by year. Playas in CRP emitted 43% less N₂O than cropland playas; grassland playas did not differ from either cropland or CRP playas (Table 1). Net emissions did not differ among landscape positions but did have a land use * year interaction. There were no differences among land use types in 2012, however, in 2013, cropland and grassland playas emitted 75% and 39% more CO_2-equiv, respectively, than playas in CRP (Table 1). Cropland and CRP playas differed among years. Cropland playa CO_2-equiv emissions increased 64% from 2012 to 2013 while emissions from playas in CRP decreased 28% from 2012 to 2013 (Table 1; Figure 1(c)). Seasonal variation existed for all GHG emissions with the bulk of emissions occurring during peak growing season.

Soil moisture differed among landscape positions and had a land use * year interaction. As expected, the wetland edge was 42% and 30% less moist than the wetland center and 1/2 distance between center and edge positions, respectively. In 2012, grassland playas had about 60% greater soil moisture than CRP and cropland playas. However, in 2013, cropland playas were 73% more moist than CRP playas; grassland playas did not differ from either cropland or CRP playas. Soil moisture in playas differed seasonally and generally declined throughout the season. Soil temperature was similar among landscape positions but had a land use * year interaction. In 2012, soil temperatures in cropland and grassland playas were 1.6˚C warmer than CRP soils. There were no differences in soil temperature among land uses in 2013. Cropland and grassland playa soils were about 0.9˚C warmer in 2012 than in 2013; soil temperatures in CRP playas did not differ between years. There was also seasonal variation in soil temperature which was highest during the summer.

Plant species richness did not differ among land uses in playas in 2013. Also, aboveground biomass did not differ among land use types in the model, however, land use type differences were evident in the Multiple Range Test. Playas in CRP had 64% greater aboveground biomass than native grassland playas; aboveground biomass in cropland playas were also 23% greater than native grassland playas but did not differ significantly from either land use type (Table 3).

There was no land use * year interaction for CO₂ emissions in uplands. Emissions of CO₂ did not differ among land use types or by year. There was a significant land use type * year interaction for CH₄ flux. There were no differences in CH₄ flux among land use types in 2012. However, in 2013, grassland uplands were a significantly stronger sink for atmospheric CH₄ than cropland or CRP uplands, which did not differ from one another (Table 1). Only grassland uplands differed between years having CH₄ sink strength on average 10 times greater in 2013 than in 2012. There was no land use * year interaction for N₂O emissions in uplands. Upland N₂O production did not differ among land use types or by year. There was also no land use * year interaction for net emissions in uplands. Net CO_2-equiv emissions did not differ among land use type or by year (Figure 1(d)). Seasonal variation existed for all GHG emissions with the bulk occurring during peak growing season.

There was a land use type * year interaction for upland soil moisture. There were no land use type differences among uplands for either year but cropland uplands were 71% more moist in 2013 than in 2012; CRP and grassland uplands did not differ between years. There was no land use type * year interaction for upland soil temperature. There were no differences in upland soil temperature among land use types or between years (Table 3).

In 2013, plant species richness differed among land uses in adjacent uplands. On average, upland native grasslands contained nearly 3 times as many plant species as croplands; uplands in CRP did not differ significantly from native grasslands or croplands (Table 3). Although there was not statistical significance, aboveground biomass in CRP uplands was nearly twice as high as cropland and native grassland uplands (Table 3).

Ecosystem restoration has been recognized in international climate policy frameworks as a possible mechanism to mitigate climate change [48]. Conservation programs and practices are widely implemented throughout the High Plains of the U.S., and with an understanding of their influence on GHG flux in the region, better-informed conservation strategies can be utilized with consideration to climate mitigation. The focus of this study was on GHG emissions from playas and watersheds, which compliments earlier studies of carbon storage dynamics throughout playas and watersheds in the High Plains [49] [50].

In playas and their watersheds, CO₂ was the greatest contributor to net emissions and because opaque collection chambers were used, samples included aboveground plant respiration as well as soil respiration. Increased deposits of upland soils and higher nutrient loads from upland runoff encourage establishment of productive, monotypic stands of vegetation that outcompete many native wetland plant species [51]. Inherently invasive plants such as Typha angustifolia (Narrowleaf Cattail) and Scirpus fluviatilis (River Bulrush) flourish under higher nutrient concentrations and produce a substantial amount of root exudates which can be an important contribution to soil C [52] and consequently higher CO₂ output. [53] documented changes in plant community composition in playas in the RWB upon restoration of former cropland and found that playas in active cropland were associated with three species of management concern; T. angustifolia, S. fluviatilis, and Phalaris arundinacea (Reed Canarygrass). The high productivity of these invasive species is likely an important factor in the higher overall CO₂ emissions from cropland playas.

Methane emissions were variable within and among playas which may be due, in part, to random ebullition events in saturated soils beneath collection chambers [54]. Emergent vegetation within chambers may also provide a route for CH₄ transport into the atmosphere. Transport of CH₄ through plant tissues has been documented in multiple studies [55] [56] [57] and is considered an important component in C cycling between wetlands and the atmosphere [58]. However, emergent vegetation may also provide an aerobic rhizosphere near the soil surface and CH₄ produced in deeper soil layers may be converted back into CO₂ by methanotrophs [59]. Although they were not statistically different, playas restored through the WRP emitted half as much CH₄, on average (error bars did not overlap), than reference condition or cropland playas which contributed to WRP having the lowest net C emissions.

Playas in WRP were drier and warmer than playas in other land uses which may be a factor in their lower overall net C emissions. Indeed, some reference condition and cropland playas received water from non-natural sources including, wetland irrigation runoff, whereas no WRP playas received supplemental water. Warmer soil temperatures in WRP playas could also be due to lower plant cover which could have increased surface evaporation rates, consequently lowering soil moisture. Sediment removal from playa basins in WRP may also result in topsoils with a lower water holding capacity due to compaction from heavy equipment and removal of lower bulk density soil organic matter overlying the high density clay pan (Bt layer) beneath playas [50].

Soil N₂O emissions are regulated by microbial nitrification/denitrification processes and in wetlands such as playas with erratic hydroperiods, N₂O emissions can be higher because wetland soils are not typically completely anoxic [21] [60]. Playas restored through the WRP had similar N₂O emissions compared to cropland and reference condition playas, suggesting playa restoration may not increase N₂O emissions. Lower aboveground plant detrital inputs due to vegetation management practices such as discing, grazing, burning, and herbicide application may limit N₂O emissions in WRP playas by limiting the amount of available NO₃ for denitrification. Removing sediments as a restoration practice in WRP playas can also remove excess N from the basin that would otherwise be available for denitrification. Indeed, [50] found that playas restored through sediment removal practices had lower soil N levels in the top 5 cm than reference condition and cropland playas.

Because there are likely lower levels of microbially available N in WRP playa soils due to sediment removal, the similarity in playa N₂O emissions among land uses may indicate changes in microbial community structure and biogeochemical functioning in WRP playas. Other studies have found land use changes and restoration activities may alter microbial communities [61] [62]. [63] found that N₂O emissions increased with plant species richness in wetland microcosms and plant diversity is often considered to be an important factor in soil microbial diversity [64] [65]. Although plant species richness did not differ among land use types in this study, [53] found that plant species richness in WRP playas was higher when sediment had been removed from basins.

Agricultural soil management activities such as fertilizer application and other cropping practices contribute 69% to U.S. N₂O emissions [17]. Croplands in the RWB receive fertilizer amendments to increase crop yields and in our study, all cropland uplands were irrigated. Excess soil N and moist soil conditions in cropland uplands made them the highest contributor to N₂O emissions, producing on average 648% more N₂O than reference condition and WRP uplands and about 450% more than playas.

Playas embedded in CRP in the WHP have 18% higher levels of C in the top 50 cm of soil than playas in cropland [49]. In our study, playas and uplands in CRP had higher aboveground biomass than playas and uplands in other land uses. However, CO₂ emissions from playas and uplands did not differ among land uses, likely due to microbial respiration being limited by soil moisture, especially in 2012. [54] found that microbial activity was inhibited when soil moisture fell below 20% VWC. On average, soil moisture in 2012 was 23% VWC in playas and 11% VWC in adjacent uplands. In 2013, rainfall totals were similar to historic averages and likely a major contributor to higher CH₄ and N₂O emissions in cropland and native grassland playas. Due to dense, exotic vegetation associated with CRP, playas embedded in CRP receive less water runoff and are inundated less frequently than playas in other land uses [26] , leading them to be smaller contributors to CH₄ emissions in 2013. Emissions of N₂O were also lower in CRP playas due to less frequent flooding, however, N₂O did not differ among land uses in uplands as they did in the RWB. In this study, only one cropland site was influenced by irrigation in the WHP; in the RWB all cropland sites were irrigated.

Native grassland uplands were the greatest atmospheric sink for CH₄ which is consistent with other studies highlighting the importance of grassland ecosystems in CH₄ oxidation [66] [67]. [68] found that microbial oxidation of CH₄ tended to be slightly higher with higher plant species diversity. Upland native grasslands had higher plant species richness than croplands which may promote CH₄ oxidation in native grassland soils in the WHP.

Our study builds on previous investigations on GHG fluxes from cropland and restored wetlands in the Prairie Pothole Region, PPR [38] [39] [40]. Within the Wetland Continuum Concept, ecological processes including those associated with GHG emissions, are driven by hydrogeomorphology and climate [69]. Because wetlands evaluated in [38] are likely similar to playas in basin morphology and water budget, differences in GHG emissions among these Great Plains regions may be primarily due to climate. Emissions of CO₂, CH₄, and N₂O were higher, on average, in RWB playas compared to PPR wetlands. It is likely that higher daily soil temperatures and extended growing seasons contributed to net emissions in RWB playas which were 34% higher than in PPR wetlands. However, wetlands in comparatively lower latitudes than the PPR, such as the RWB, also have higher evapotranspiration rates and regions receiving less annual rainfall such as the WHP, warmer soils may not lead to higher GHG emissions due to soil moisture restrictions. Emissions from WHP playas were about 33% and 50% lower than PPR and RWB wetlands, respectively. Other localized factors including soils, floral/faunal composition, and management practices could also have contributed to regional differences in emissions.

Although wetlands continue to be degraded and an estimated 50% of all wetlands have been lost worldwide, the services afforded by these ecosystems are highly valued [70]. Wetland restoration efforts and “no net loss” policy goals are in place to prevent further wetland loss and restore ecosystem services in the U.S. However, in some cases, such efforts may have a negative influence on climate [6] [16].

In the RWB, wetland restorations through the WRP have restored about 2600 ha of wetlands and uplands in a region that has lost 90% of its historic wetland area [36] [71]. A primary goal for wetland restoration in the RWB is to provide adequate forage habitat for migrating waterfowl [72]. Managing vegetation in playas in WRP to prevent dense perennials such as T. angustifolia, S. fluviatilis, and P. arundinacea through discing, grazing, burning, and herbicide application likely limits the spread of these invasives resulting in lower aboveground plant biomass and a higher diversity of plant species [53]. Restoration practices such as sediment removal can also change plant species composition as well as lower excess nutrients in playas [53] [73]. Practices associated with WRP in the RWB effectively reduce GHG emissions from playas, thus climate mitigation is an ecosystem service that may be enhanced through playa restoration in the RWB.

Due to its extent in the High Plains, CRP has a greater influence on playas than any other conservation program [19]. Playas embedded in CRP produced less net CO_2-equiv emissions than native grassland and cropland playas. Because they are inundated less frequently as a result of exotic grasses in the upland, however, their capacity to provide other services such as wildlife habitat and aquifer recharge may be reduced [19]. The benefits of lower gas emissions must be appropriately weighted against loss of services related to shorter hydroperiods in playas in CRP. Furthermore, native grassland watersheds in the WHP should be maintained in order to maximize atmospheric CH₄ oxidation and climate mitigation services.