Geology Reports⌕ Search

SEARCH · Geology Reports

Results for “Agricultural and Forest Meteorology”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2Linked to original sources
USGS2014

Net ecosystem productivity of temperate grasslands in northern China: An upscaling study

Grassland is one of the widespread biome types globally, and plays an important role in the terrestrial carbon cycle. We examined net ecosystem production (NEP) for the temperate grasslands in northern China from 2000 to 2010. We combined flux observations, satellite data, and climate data to develop a piecewise regression model for NEP, and then used the model to map NEP for grasslands in northern China. Over the growing season, the northern China's grassland had a net carbon uptake of 158 ± 25 g C m −2 during 2000–2010 with the mean regional NEP estimate of 126 Tg C. Our results showed generally higher grassland NEP at high latitudes (northeast) than at low latitudes (central and west) because of different grassland types and environmental conditions. In the northeast, which is dominated by meadow steppes, the growing season NEP generally reached 200–300 g C m −2 . In the southwest corner of the region, which is partially occupied by alpine meadow systems, the growing season NEP also reached 200–300 g C m −2 . In the central part, which is dominated by typical steppe systems, the growing season NEP generally varied in the range of 100–200 g C m−2. The NEP of the northern China's grasslands was highly variable through years, ranging from 129 (2001) to 217 g C m −2 growing season −1 (2010). The large interannual variations of NEP could be attributed to the sensitivity of temperate grasslands to climate changes and extreme climatic events. The droughts in 2000, 2001, and 2006 reduced the carbon uptake over the growing season by 11%, 29%, and 16% relative to the long-term (2000–2010) mean. Over the study period (2000–2010), precipitation was significantly correlated with NEP for the growing season (R 2 = 0.35, p-value < 0.1), indicating that water availability is an important stressor for the productivity of the temperate grasslands in semi-arid and arid regions in northern China. We conclude that northern temperate grasslands have the potential to sequester carbon, but the capacity of carbon sequestration depends on grassland types and environmental conditions. Extreme climate events like drought can significantly reduce the net carbon uptake of grasslands.

Inner Mongolia;Gansu;And Ningxia Provinces↗
USGS2015

Bistability of mangrove forests and competition with freshwater plants

Halophytic communities such as mangrove forests and buttonwood hammocks tend to border freshwater plant communities as sharp ecotones. Most studies attribute this purely to underlying physical templates, such as groundwater salinity gradients caused by tidal flux and topography. However, a few recent studies hypothesize that self-reinforcing feedback between vegetation and vadose zone salinity are also involved and create a bistable situation in which either halophytic dominated habitat or freshwater plant communities may dominate as alternative stable states. Here, we revisit the bistability hypothesis and demonstrate the mechanisms that result in bistability. We demonstrate with remote sensing imagery the sharp boundaries between freshwater hardwood hammock communities in southern Florida and halophytic communities such as buttonwood hammocks and mangroves. We further document from the literature how transpiration of mangroves and freshwater plants respond differently to vadose zone salinity, thus altering the salinity through feedback. Using mathematical models, we show how the self-reinforcing feedback, together with physical template, controls the ecotones between halophytic and freshwater communities. Regions of bistability along environmental gradients of salinity have the potential for large-scale vegetation shifts following pulse disturbances such as hurricane tidal surges in Florida, or tsunamis in other regions. The size of the region of bistability can be large for low-lying coastal habitat due to the saline water table, which extends inland due to salinity intrusion. We suggest coupling ecological and hydrologic processes as a framework for future studies.

Florida↗
USGS2015

Landscape-level terrestrial methane flux observed from a very tall tower

Simulating the magnitude and variability of terrestrial methane sources and sinks poses a challenge to ecosystem models because the biophysical and biogeochemical processes that lead to methane emissions from terrestrial and freshwater ecosystems are, by their nature, episodic and spatially disjunct. As a consequence, model predictions of regional methane emissions based on field campaigns from short eddy covariance towers or static chambers have large uncertainties, because measurements focused on a particular known source of methane emission will be biased compared to regional estimates with regards to magnitude, spatial scale, or frequency of these emissions. Given the relatively large importance of predicting future terrestrial methane fluxes for constraining future atmospheric methane growth rates, a clear need exists to reduce spatiotemporal uncertainties. In 2010, an Ameriflux tower (US-PFa) near Park Falls, WI, USA, was instrumented with closed-path methane flux measurements at 122 m above ground in a mixed wetland&ndash;upland landscape representative of the Great Lakes region. Two years of flux observations revealed an average annual methane (CH 4 ) efflux of 785 &plusmn; 75 mg C CH 4 m &minus;2 yr &minus;1 , compared to a mean CO 2 sink of &minus;80 g C CO 2 m &minus;2 yr &minus;1 , a ratio of 1% in magnitude on a mole basis. Interannual variability in methane flux was 30% of the mean flux and driven by suppression of methane emissions during dry conditions in late summer 2012. Though relatively small, the magnitude of the methane source from the very tall tower measurements was mostly within the range previously measured using static chambers at nearby wetlands, but larger than a simple scaling of those fluxes to the tower footprint. Seasonal patterns in methane fluxes were similar to those simulated in the Dynamic Land Ecosystem Model (DLEM), but magnitude depends on model parameterization and input data, especially regarding wetland extent. The model was unable to simulate short-term (sub-weekly) variability. Temperature was found to be a stronger driver of regional CH 4 flux than moisture availability or net ecosystem production at the daily to monthly scale. Taken together, these results emphasize the multi-timescale dependence of drivers of regional methane flux and the importance of long, continuous time series for their characterization.

Wisconsin↗
USGS2014

Approximations of stand water use versus evapotranspiration from three mangrove forests in southwest Florida, USA

Leaves from mangrove forests are often considered efficient in the use of water during photosynthesis, but less is known about whole-tree and stand-level water use strategies. Are mangrove forests as conservative in water use as experimental studies on seedlings imply? Here, we apply a simple model to estimate stand water use ( S ), determine the contribution of S to evapotranspiration (ET), and approximate the distribution of S versus ET over annual cycles for three mangrove forests in southwest Florida, USA. The value of S ranged from 350 to 511 mm year &minus;1 for two mangrove forests in Rookery Bay to 872 mm year &minus;1 for a mangrove forest along the Shark River in Everglades National Park. This represents 34&ndash;49% of ET for Rookery Bay mangroves, a rather conservative rate of S , and 63&ndash;66% of ET for the Shark River mangroves, a less conservative rate of S . However, variability in estimates of S in mangroves is high enough to require additional study on the spatial changes related to forest structural shifts, different tidal regimes, and variable site-specific salinity concentrations in multiple mangrove forests before a true account of water use conservation strategies can be understood at the landscape scale. Evidence does suggest that large, well-developed mangrove forests have the potential to contribute considerably to the ET balance; however, regionally most mangrove forests are much smaller in stature in Florida and likely contribute less to regional water losses through stand-level transpiration.

Florida↗
USGS2015

Component-specific dynamics of riverine mangrove CO 2 efflux in the Florida coastal Everglades

Carbon cycling in mangrove forests represents a significant portion of the coastal wetland carbon (C) budget across the latitudes of the tropics and subtropics. Previous research suggests fluctuations in tidal inundation, temperature and salinity can influence forest metabolism and C cycling. Carbon dioxide (CO2) from respiration that occurs from below the canopy is contributed from different components. In this study, we investigated variation in CO2 flux among different below-canopy components (soil, leaf litter, course woody debris, soil including pneumatophores, prop roots, and surface water) in a riverine mangrove forest of Shark River Slough estuary, Everglades National Park (Florida, USA). The range in CO2 flux from different components exceeded that measured among sites along the oligohaline-saline gradient. Black mangrove (Avicennia germinans) pneumatophores contributed the largest average CO2 flux. Over a narrow range of estuarine salinity (25&ndash;35 practical salinity units (PSU)), increased salinity resulted in lower CO2 flux to the atmosphere. Tidal inundation reduced soil CO2 flux overall but increased the partial pressure of CO2 (pCO2) observed in the overlying surface water upon flooding. Higher pCO2 in surface water is then subject to tidally driven export, largely as HCO3. Integration and scaling of CO2 flux rates to forest scale allowed for improved understanding of the relative contribution of different below-canopy components to mangrove forest ecosystem respiration (ER). Summing component CO2fluxes suggests a more significant contribution of below-canopy respiration to ER than previously considered. An understanding of below-canopy CO2 component fluxes and their contributions to ER can help to elucidate how C cycling will change with discrete disturbance events (e.g., hurricanes) and long-term change, including sea-level rise, and potential impact mangrove forests. As such, key controls on below-canopy ER must be taken into consideration when developing and modeling mangrove forest C budgets.

Florida↗
USGS2016

Hyperspectral narrowband and multispectral broadband indices for remote sensing of crop evapotranspiration and its components (transpiration and soil evaporation)

Evapotranspiration (ET) is an important component of micro- and macro-scale climatic processes. In agriculture, estimates of ET are frequently used to monitor droughts, schedule irrigation, and assess crop water productivity over large areas. Currently, in situ measurements of ET are difficult to scale up for regional applications, so remote sensing technology has been increasingly used to estimate crop ET. Ratio-based vegetation indices retrieved from optical remote sensing, like the Normalized Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index, and Enhanced Vegetation Index are critical components of these models, particularly for the partitioning of ET into transpiration and soil evaporation. These indices have their limitations, however, and can induce large model bias and error. In this study, micrometeorological and spectroradiometric data collected over two growing seasons in cotton, maize, and rice fields in the Central Valley of California were used to identify spectral wavelengths from 428 to 2295 nm that produced the highest correlation to and lowest error with ET, transpiration, and soil evaporation. The analysis was performed with hyperspectral narrowbands (HNBs) at 10 nm intervals and multispectral broadbands (MSBBs) commonly retrieved by Earth observation platforms. The study revealed that (1) HNB indices consistently explained more variability in ET (&Delta; R 2 = 0.12), transpiration (&Delta; R 2 = 0.17), and soil evaporation (&Delta; R 2 = 0.14) than MSBB indices; (2) the relationship between transpiration using the ratio-based index most commonly used for ET modeling, NDVI, was strong ( R 2 = 0.51), but the hyperspectral equivalent was superior ( R 2 = 0.68); and (3) soil evaporation was not estimated well using ratio-based indices from the literature (highest R 2 = 0.37), but could be after further evaluation, using ratio-based indices centered on 743 and 953 nm ( R 2 = 0.72) or 428 and 1518 nm ( R 2 = 0.69).

California↗
USGS2017

A method for quantifying cloud immersion in a tropical mountain forest using time-lapse photography

Quantifying the frequency, duration, and elevation range of fog or cloud immersion is essential to estimate cloud water deposition in water budgets and to understand the ecohydrology of cloud forests. The goal of this study was to develop a low-cost and high spatial-coverage method to detect occurrence of cloud immersion within a mountain cloud forest by using time-lapse photography. Trail cameras and temperature/relative humidity sensors were deployed at five sites covering the elevation range from the assumed lifting condensation level to the mountain peaks in the Luquillo Mountains of Puerto Rico. Cloud-sensitive image characteristics (contrast, the coefficient of variation and the entropy of pixel luminance, and image colorfulness) were used with a k -means clustering approach to accurately detect cloud-immersed conditions in a time series of images from March 2014 to May 2016. Images provided hydrologically meaningful cloud-immersion information while temperature-relative humidity data were used to refine the image analysis using dew point information and provided temperature gradients along the elevation transect. Validation of the image processing method with human-judgment based classification generally indicated greater than 90% accuracy. Cloud-immersion frequency averaged 80% at sites above 900 m during nighttime hours and 49% during daytime hours, and was consistent with diurnal patterns of cloud immersion measured in a previous study. Results for the 617 m site demonstrated that cloud immersion in the Luquillo Mountains rarely occurs at the previously-reported cloud base elevation of about 600 m (11% during nighttime hours and 5% during daytime hours). The framework presented in this paper will be used to monitor at a low cost and high spatial resolution the long-term variability of cloud-immersion patterns in the Luquillo Mountains, and can be applied to ecohydrology research at other cloud-forest sites or in coastal ecosystems with advective sea fog.

El Yunque National Forest, Puerto Rico↗
USGS2018

Using a Bayesian framework to account for advection in seven years of snowpack CO2 fluxes in a mortality-impacted subalpine forest

In subalpine forests where snowpacks can persist for 50% of the year or longer, wintertime subnivean soil respiration constitutes a significant, but poorly understood, portion of the carbon (C) cycle. The temporal and seasonal dynamics of this flux are difficult to measure compared to snow-free respiration. This makes it challenging to fully address key knowledge gaps in high-elevation forests, including effects of tree mortality on ecosystem C fluxes. Our paper attempts to overcome these challenges by developing a new method for estimating wintertime CO 2 flux from soil underneath winter snowpack based on CO 2 profiles in the snow and using this to examine potential effects of a tree mortality event on wintertime respiration. The chief problem for estimates based on snow CO 2 measurements is that during strong winds, vertical CO 2 profiles can depart significantly from linearity, resulting in CO 2 fluxes that are not steady-state diffusive fluxes. Consequently, these fluxes include an advective component induced by wind and atmospheric pressure, resulting in above-snowpack fluxes that may be temporally decoupled from biological flux production in the soil. These advective fluxes lead to errors in estimates of the total amount of CO 2 produced from the soil surface during the winter, and may complicate the detection of significant seasonal and interannual trends in snowpack CO 2 fluxes. Our goal was twofold: first, to improve methods for estimating wintertime subnivean soil respiration; and second, to use this new method to examine potential effects of forest mortality on respiration. We develop a Bayesian-based model of CO 2 snowpack profiles and accounts for CO 2 storage flux within the snowpack to calculate diffusive and advective fluxes from both the bottom and the top of the snowpack. We apply this model to hourly snowpack CO 2 profile data collected over 7 years at an AmeriFlux eddy covariance (EC) site in a Wyoming subalpine forest that had recently experienced a mass tree mortality event. Hourly advective fluxes from the snowpack surface are less variable over time than hourly EC fluxes. Nonetheless, on a daily time scale, the two methods agree well with each other. We found that soil biological CO 2 production entering the bottom of snowpack peaked 3 years after the onset of the bark beetle outbreak and 1-year after dead trees dropped their needles, presenting the first evidence of a short-term increase in respiration following bark beetle mortality. Our findings emphasize the importance of accounting for the role of wind-driven advection in wintertime respiration and suggest that tree mortality may nearly double wintertime respiration in the first few years.

Wyoming↗
USGS2018

Shrubland carbon sink depends upon winter water availability in the warm deserts of North America

Global-scale studies suggest that dryland ecosystems dominate an increasing trend in the magnitude and interannual variability of the land CO 2 sink. However, such model-based analyses are poorly constrained by measured CO 2 exchange in open shrublands, which is the most common global land cover type, covering ∼14% of Earth’s surface. Here we evaluate how the amount and seasonal timing of water availability regulate CO 2 exchange between shrublands and the atmosphere. We use eddy covariance data from six US sites across the three warm deserts of North America with observed ranges in annual precipitation of ∼100–400mm, annual temperatures of 13–18°C, and records of 2–8 years (33 site-years in total). The Chihuahuan, Sonoran and Mojave Deserts present gradients in both mean annual precipitation and its seasonal distribution between the wet-winter Mojave Desert and the wet-summer Chihuahuan Desert. We found that due to hydrologic losses during the wettest summers in the Sonoran and Chihuahuan Deserts, evapotranspiration (ET) was a better metric than precipitation of water available to drive dryland CO 2 exchange. In contrast with recent synthesis studies across diverse dryland biomes, we found that NEP could not be directly predicted from ET due to wintertime decoupling of the relationship between ecosystem respiration (R eco ) and gross ecosystem productivity (GEP). Ecosystem water use efficiency (WUE=GEP/ET) did not differ between winter and summer. Carbon use efficiency (CUE=NEP/GEP), however, was greater in winter because R eco returned a smaller fraction of carbon to the atmosphere (23% of GEP) than in summer (77%). Combining the water-carbon relations found here with historical precipitation since 1980, we estimate that lower average winter precipitation during the 21st century reduced the net carbon sink of the three deserts by an average of 6.8TgC yr 1 . Our results highlight that winter precipitation is critical to the annual carbon balance of these warm desert shrublands.

Agricultural and Forest Meteorology↗
USGS2018

On the exchange of sensible and latent heat between the atmosphere and melting snow

The snow energy balance is difficult to measure during the snowmelt period, yet critical for predictions of water yield in regions characterized by snow cover. Robust simplifications of the snowmelt energy balance can aid our understanding of water resources in a changing climate. Research to date has demonstrated that the net turbulent flux ( F T ) between a melting snowpack and the atmosphere is negligible if the sum of atmospheric vapor pressure ( e a ) and temperature ( T a ) equals a constant, but it is unclear how frequently this situation holds across different sites. Here, we quantified the contribution of F T to the snowpack energy balance during 59 snowmelt periods across 11 sites in the FLUXNET2015 database with a detailed analysis of snowmelt in subarctic tundra near Abisko, Sweden. At the Abisko site we investigated the frequency of occurrences during which sensible heat flux ( H ) and latent heat flux ( λE ) are of (approximately) equal but opposite sign, and if the sum of these terms, F T , is therefore negligible during the snowmelt period. H approximately equaled - λE for less than 50% of the melt period and F T was infrequently a trivial term in the snowmelt energy balance at Abisko. The reason is that the relationship between observed e a and T a is roughly orthogonal to the “line of equality” at which H equals - λE as warmer T a during the melt period usually resulted in greater e a . This relationship holds both within melt periods at individual sites and across different sites in the FLUXNET2015 database, where F T comprised less than 20% of the energy available to melt snow, Q m , in 44% of the snowmelt periods studied here. F T / Q m was significantly related to the mean e a during the melt period, but not mean T a , and F T tended to be near 0 W m −2 when e a averaged ca . 0.5 kPa. F T may become an increasingly important term in the snowmelt energy balance across many global regions as warmer temperatures are projected to cause snow to melt more slowly and earlier in the year under conditions of lower net radiation ( R n ). Eddy covariance research networks such as Ameriflux must improve their ability to observe cold-season processes to enhance our understanding of water resources and surface-atmosphere exchange in a changing climate.

Agricultural and Forest Meteorology↗
USGS2019

Time to branch out? Application of hierarchical survival models in plant phenology

The sensitivity of phenology to environmental drivers can vary across geography and species. As such, models developed to predict phenology are typically site- or taxon-specific. Generation of site- and taxon-specific models is limited by the intensive in-situ phenological monitoring effort required to generate sufficient data to parameterize each model. Where in-situ phenological observations exist, the data are often subject to analytical issues due to the limited duration of any individual monitoring program, spotty site- and species- level coverage, lack of standardized methodology, and infrequent or variable census intervals. Together, these characteristics constrain our ability to make phenological inferences outside of select sites and taxa where long-duration, intensive monitoring has occurred. In this study, we leveraged two national, standardized phenology datasets to develop a multi-species and multi-site state-space survival model of the onset of deciduous tree and shrub spring (leaf out) and fall (leaf-color) events across temperate ecoregions of the United States. We used data from two national-scale phenological databases, a 9-year, broadly distributed dataset from the USA National Phenology Network and a 4-year dataset from the National Ecological Observatory Network, to quantify regional and interspecific variation in sensitivity to environmental drivers for both spring and fall leaf phenophases. Spring leaf out was generally promoted by longer days, spring growing degree day accumulation, overwinter chilling, and was suppressed by frost events, whereas fall leaf color was promoted by shorter days and cold accumulation. The sensitivity to most environmental drivers tended to be more variable among species than among the regions as defined here (EPA ecoregions of North America, excluding desert and tropical areas). The results of this study lay the groundwork for incorporating the growing collection of phenological observations into a generalized framework for predicting the transition states for any species, in any location.

Agricultural and Forest Meteorology↗
USGS2021

Performance of the ecosystem demography model (EDv2.2) in simulating gross primary production capacity and activity in a dryland study area

Dryland ecosystems play an important role in the global carbon cycle, including regulating the inter-annual global carbon sink. Dynamic global vegetation models (DGVMs) are essential tools that can help us better understand carbon cycling in different ecosystems. Currently, there is limited knowledge of the performance of these models in drylands partly due to characterizing the heterogeneity of the vegetation and hydrometeorological conditions. The aim of this study is to evaluate the performance of a DGVM for drylands to facilitate improved understanding of gross primary production (GPP) as one of the important components of the carbon cycle. We performed a sensitivity analysis and calibrated the Ecosystem Demography (EDv2.2) DGVM to simulate GPP in a dryland watershed (Reynolds Creek Experimental Watershed, Idaho) in the western US for the years 2000-2017. GPP capacity and activity were investigated by comparing model simulations with GPP estimated from eddy covariance data (available from 2015-2017) and remote sensing products (2000-2017). Our results show good performance of EDv2.2 at daily timesteps ( R M S E &#x2248; 0.38 [ kgC / m 2 / year ] ) "> RMSE≈0.38[kgC/m2/year])

Idaho↗
USGS2021

Only sun-lit leaves of the uppermost canopy exceed both air temperature and photosynthetic thermal optima in a wet tropical forest

Tropical forests have evolved under relatively narrow temperature regimes, and therefore may be more susceptible to climatic change than forests in higher latitudes. Recent evidence shows that lowland tropical forest canopies may already be exceeding thermal maxima for photosynthesis. Height can strongly influence both the microclimate and physiology of forest canopy foliage, yet vertical trends in canopy micrometeorology are rarely examined in tropical forests. To improve our understanding of how climatological and micrometeorological conditions affect tropical tree function, we assessed vertical gradients of photosynthetic photon flux density, vapor pressure deficit, air temperature, leaf temperature, and the difference between leaf and air temperature (ΔT) in a Puerto Rican tropical wet forest. Both air temperature and vapor pressure deficit increased linearly with height. Leaf temperature, however, did not significantly differ across the shaded foliage from 0-16 m, while the uppermost layer (20 m) was up to 4°C hotter than the rest of the foliage and up to 5°C hotter than air temperature at the highest radiation intensity. As a result, leaf temperatures in the shaded middle canopy and understory showed nearly poikilothermic behavior (i.e., leaf temperatures = air temperature), while the uppermost canopy strata showed megathermic behavior (i.e., leaf temperatures greater than air temperature), revealing different thermoregulation strategies for sun-lit versus shaded foliage. In addition, the uppermost canopy was the only stratum to exceed mean photosynthetic temperature optima for this site (T opt = 30.2 ± 1.1°C). Because the upper canopy plays a disproportionately large role in whole-forest photosynthesis, continued warming could potentially weaken the tropics’ carbon sink capacity. However, the shaded leaves may be able increase carbon uptake with further warming because they appear to be able to maintain temperatures below photosynthetic optima, possibly with the help of radiation shielding provided by the uppermost canopy layer.

Agricultural and Forest Meteorology↗
USGS2021

Forest evapotranspiration dynamics over a fragmented forest landscape under drought in southwestern Amazonia

Ongoing climate change and human conversion of forests to other land uses alter regional evapotranspiration dynamics and, consequently, impact associated hydrological systems in Amazonia. We studied the effects of drought and fragmentation on forest evapotranspiration using the surface energy balance-based model METRIC (Mapping Evapotranspiration at high Resolution with Internalized Calibration) for a fragmented forest landscape in Brazil's Amazonian state of Rondônia. Dry season (June-August) forest evapotranspiration estimates were produced for the 2009-2011 period that encompassed the 2010 drought event, one of the extreme droughts in the Amazon. METRIC evapotranspiration data were analyzed in relation to climate (monthly precipitation and cumulative water deficit) and forest fragmentation (edge distance from 100m to 1000m from forest edge and edge density). During the dry season of 2009, pre-drought, forest evapotranspiration did not fall below 110mm/month. However, the 2010 drought year showed a drastic decline in evapotranspiration by 32%, to 75mm/month, from July to August. In 2011, evapotranspiration rates were still depressed with August rates dropping as low as 85mm/month. Forest evapotranspiration dynamics were driven mainly by precipitation and corresponding water deficits in the drier years (2010 and 2011), although evapotranspiration deficits along the edges of forest fragments were locally significant, at the landscape scale. The forests near edges (to 100m) had progressively lower evapotranspiration levels than interior forests as dry seasons progressed and these differences were greatest in the 2010 drought year, reaching almost 5%. Our results suggest that during the driest months, fragmentation exacerbated both the rate and extent of evapotranspiration reductions over forest areas up to 100m from edges, equivalent to ~20% of the forested landscape in our study area.

Rondonia↗
USGS2021

Gap-filling eddy covariance methane fluxes: Comparison of machine learning model predictions and uncertainties at FLUXNET-CH4 wetlands

Time series of wetland methane fluxes measured by eddy covariance require gap-filling to estimate daily, seasonal, and annual emissions. Gap-filling methane fluxes is challenging because of high variability and complex responses to multiple drivers. To date, there is no widely established gap-filling standard for wetland methane fluxes, with regards both to the best model algorithms and predictors. This study synthesizes results of different gap-filling methods systematically applied at 17 wetland sites spanning boreal to tropical regions and including all major wetland classes and two rice paddies. Procedures are proposed for: 1) creating realistic artificial gap scenarios, 2) training and evaluating gap-filling models without overstating performance, and 3) predicting half-hourly methane fluxes and annual emissions with realistic uncertainty estimates. Performance is compared between a conventional method (marginal distribution sampling) and four machine learning algorithms. The conventional method achieved similar median performance as the machine learning models but was worse than the best machine learning models and relatively insensitive to predictor choices. Of the machine learning models, decision tree algorithms performed the best in cross-validation experiments, even with a baseline predictor set, and artificial neural networks showed comparable performance when using all predictors. Soil temperature was frequently the most important predictor whilst water table depth was important at sites with substantial water table fluctuations, highlighting the value of data on wetland soil conditions. Raw gap-filling uncertainties from the machine learning models were underestimated and we propose a method to calibrate uncertainties to observations. The python code for model development, evaluation, and uncertainty estimation is publicly available. This study outlines a modular and robust machine learning workflow and makes recommendations for, and evaluates an improved baseline of, methane gap-filling models that can be implemented in multi-site syntheses or standardized products from regional and global flux networks (e.g., FLUXNET).

Agricultural and Forest Meteorology↗
USGS2022

From flowering to foliage: Accelerometers track tree sway to provide high-resolution insights into tree phenology

Trees are bioindicators of global climate change and regional urbanization, but available monitoring tools are ineffective for fine-scale observation of many species. Using six accelerometers mounted on two urban ash trees ( Fraxinus americana ), we looked at high-frequency tree vibrations, or change in periodicity of tree sway as a proxy for mass changes, to infer seasonal patterns of flowering and foliage (phenophases). We compared accelerometer-estimated phenophases to those derived from digital repeat photography using Green Chromatic Coordinates (GCC) and visual observation of phenophases defined by the USA National Phenology Network (NPN). We also drew comparisons between two commercial accelerometers and assessed how placement height influenced the ability to extract seasonal transition dates. Most notably, tree sway data showed a greenness signal in an urban environment and produced a clear flowering time-series and peak flowering signal (PF), marking the first observations of a flower phenophase using accelerometer data. Estimated start of spring (SOS) from accelerometers and time-lapse GCC were more similar than start of autumn (SOA); accelerometers lagged behind the time-lapse camera dates by three and four days for SOS and 13 and 14 days for SOA for each tree. Estimates for SOS and SOA from accelerometers and time-lapse cameras aligned closely with different NPN phenophases. The two commercial accelerometers produced similar season onset: a difference of 2.4 to 3.8 days for SOS, 2.1 days for SOA, and 0.5 to 2.0 days for PF. Accelerometers placed at the main crown branch point versus higher in the canopy showed a difference of 0.2 to 4.9 days for SOS and -1.5 to 1.7 days for PF. Our results suggest accelerometers present a novel opportunity to objectively monitor reproductive tree biology and fill gaps in phenology observations. Furthermore, widely available accelerometers show promise for scaling up from individual trees to the landscape level to aid forest management and assessing climate change impacts to tree phenology.

Colorado↗
USGS2022

Prairie wetlands as sources or sinks of nitrous oxide: Effects of land use and hydrology

National and global greenhouse gas (GHG) budgets are continually being refined as data become available. Primary sources of the potent GHG nitrous oxide (N 2 O) include agricultural soil management and burning of fossil fuels, but comprehensive N 2 O budgets also incorporate less prominent factors such as wetlands. Freshwater wetland GHG flux estimates, however, have high uncertainty, and wetlands have been identified as both sources and sinks. Here, we analyzed a regional database of >26,000 N 2 O chamber flux measurements sampled across >150 wetlands from the Prairie Pothole Region (PPR) in the Great Plains of North America. Our goal was to identify important land use and hydrologic drivers of N 2 O flux to help reduce uncertainty in N 2 O models, and to incorporate these drivers into an upscaled estimate of wetland N 2 O emissions from the U.S. portion of the PPR. Within individual wetlands, exposed soils with no standing water, such as along wetland edges, were hotspots that accounted for greater than 90% of wetland N 2 O emissions. In contrast wet (i.e., ponded) areas had minimal or negative N 2 O flux. N 2 O flux from wetlands nested within croplands (16.3–17.3 μg N 2 O m −2 hr −1 ) was, in some instances, nearly double that from wetlands within grasslands (9.2–14.4 μg N 2 O m −2 h −1 ). We estimated that seasonal N 2 O flux from PPR wetlands equated to roughly 0.2% (1.04 Tg CO 2 equivalents) of the U.S. N 2 O budget (c. 2019). Overall, even though PPR wetlands are a small net source of N 2 O to the atmosphere, their emissions are negligible relative to agricultural soil management. Policy and management to restore wetland hydrology and surrounding uplands from cropland to grasslands can reduce landscape N 2 O fluxes. Future activities focused on wetland N 2 O flux would benefit from inclusion of adjacent land use and hydrologic factors, as well as from incorporation of temporally dynamic ponded wetland areas.

Iowa, Minnesota, Montana, North Dakota, South Dako↗