A global land cover training dataset from 1984 to 2020.

State-of-the-art cloud computing platforms such as Google Earth Engine (GEE) enable regional-to-global land cover and land cover change mapping with machine learning algorithms. However, collection of high-quality training data, which is necessary for accurate land cover mapping, remains costly and labor-intensive. To address this need, we created a global database of nearly 2 million training units spanning the period from 1984 to 2020 for seven primary and nine secondary land cover classes. Our training data collection approach leveraged GEE and machine learning algorithms to ensure data quality and biogeographic representation. We sampled the spectral-temporal feature space from Landsat imagery to efficiently allocate training data across global ecoregions and incorporated publicly available and collaborator-provided datasets to our database. To reflect the underlying regional class distribution and post-disturbance landscapes, we strategically augmented the database. We used a machine learning-based cross-validation procedure to remove potentially mis-labeled training units. Our training database is relevant for a wide array of studies such as land cover change, agriculture, forestry, hydrology, urban development, among many others.

A high spatial resolution land surface phenology dataset for AmeriFlux and NEON sites.

Vegetation phenology is a key control on water, energy, and carbon fluxes in terrestrial ecosystems. Because vegetation canopies are heterogeneous, spatially explicit information related to seasonality in vegetation activity provides valuable information for studies that use eddy covariance measurements to study ecosystem function and land-atmosphere interactions. Here we present a land surface phenology (LSP) dataset derived at 3 m spatial resolution from PlanetScope imagery across a range of plant functional types and climates in North America. The dataset provides spatially explicit information related to the timing of phenophase changes such as the start, peak, and end of vegetation activity, along with vegetation index metrics and associated quality assurance flags for the growing seasons of 2017-2021 for 10 × 10 km windows centred over 104 eddy covariance towers at AmeriFlux and National Ecological Observatory Network (NEON) sites. These LSP data can be used to analyse processes controlling the seasonality of ecosystem-scale carbon, water, and energy fluxes, to evaluate predictions from land surface models, and to assess satellite-based LSP products.

Extensive land cover change across Arctic-Boreal Northwestern North America from disturbance and climate forcing.

A multitude of disturbance agents, such as wildfires, land use, and climate-driven expansion of woody shrubs, is transforming the distribution of plant functional types across Arctic-Boreal ecosystems, which has significant implications for interactions and feedbacks between terrestrial ecosystems and climate in the northern high-latitude. However, because the spatial resolution of existing land cover datasets is too coarse, large-scale land cover changes in the Arctic-Boreal region (ABR) have been poorly characterized. Here, we use 31 years (1984-2014) of moderate spatial resolution (30 m) satellite imagery over a region spanning 4.7 × 106  km2 in Alaska and northwestern Canada to characterize regional-scale ABR land cover changes. We find that 13.6 ± 1.3% of the domain has changed, primarily via two major modes of transformation: (a) simultaneous disturbance-driven decreases in Evergreen Forest area (-14.7 ± 3.0% relative to 1984) and increases in Deciduous Forest area (+14.8 ± 5.2%) in the Boreal biome; and (b) climate-driven expansion of Herbaceous and Shrub vegetation (+7.4 ± 2.0%) in the Arctic biome. By using time series of 30 m imagery, we characterize dynamics in forest and shrub cover occurring at relatively short spatial scales (hundreds of meters) due to fires, harvest, and climate-induced growth that are not observable in coarse spatial resolution (e.g., 500 m or greater pixel size) imagery. Wildfires caused most of Evergreen Forest Loss and Evergreen Forest Gain and substantial areas of Deciduous Forest Gain. Extensive shifts in the distribution of plant functional types at multiple spatial scales are consistent with observations of increased atmospheric CO2 seasonality and ecosystem productivity at northern high-latitudes and signal continental-scale shifts in the structure and function of northern high-latitude ecosystems in response to climate change.

Publisher Correction: Tracking vegetation phenology across diverse biomes using Version 2.0 of the PhenoCam Dataset.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Tracking vegetation phenology across diverse biomes using Version 2.0 of the PhenoCam Dataset.

Monitoring vegetation phenology is critical for quantifying climate change impacts on ecosystems. We present an extensive dataset of 1783 site-years of phenological data derived from PhenoCam network imagery from 393 digital cameras, situated from tropics to tundra across a wide range of plant functional types, biomes, and climates. Most cameras are located in North America. Every half hour, cameras upload images to the PhenoCam server. Images are displayed in near-real time and provisional data products, including timeseries of the Green Chromatic Coordinate (Gcc), are made publicly available through the project web page ( https://phenocam.sr.unh.edu/webcam/gallery/ ). Processing is conducted separately for each plant functional type in the camera field of view. The PhenoCam Dataset v2.0, described here, has been fully processed and curated, including outlier detection and expert inspection, to ensure high quality data. This dataset can be used to validate satellite data products, to evaluate predictions of land surface models, to interpret the seasonality of ecosystem-scale CO2 and H2O flux data, and to study climate change impacts on the terrestrial biosphere.

Tracking vegetation phenology across diverse North American biomes using PhenoCam imagery.

Vegetation phenology controls the seasonality of many ecosystem processes, as well as numerous biosphere-atmosphere feedbacks. Phenology is also highly sensitive to climate change and variability. Here we present a series of datasets, together consisting of almost 750 years of observations, characterizing vegetation phenology in diverse ecosystems across North America. Our data are derived from conventional, visible-wavelength, automated digital camera imagery collected through the PhenoCam network. For each archived image, we extracted RGB (red, green, blue) colour channel information, with means and other statistics calculated across a region-of-interest (ROI) delineating a specific vegetation type. From the high-frequency (typically, 30 min) imagery, we derived time series characterizing vegetation colour, including "canopy greenness", processed to 1- and 3-day intervals. For ecosystems with one or more annual cycles of vegetation activity, we provide estimates, with uncertainties, for the start of the "greenness rising" and end of the "greenness falling" stages. The database can be used for phenological model validation and development, evaluation of satellite remote sensing data products, benchmarking earth system models, and studies of climate change impacts on terrestrial ecosystems.

Accounting for urban biogenic fluxes in regional carbon budgets.

Many ecosystem models incorrectly treat urban areas as devoid of vegetation and biogenic carbon (C) fluxes. We sought to improve estimates of urban biomass and biogenic C fluxes using existing, nationally available data products. We characterized biogenic influence on urban C cycling throughout Massachusetts, USA using an ecosystem model that integrates improved representation of urban vegetation, growing conditions associated with urban heat island (UHI), and altered urban phenology. Boston's biomass density is 1/4 that of rural forests, however 87% of Massachusetts' urban landscape is vegetated. Model results suggest that, kilogram-for-kilogram, urban vegetation cycles C twice as fast as rural forests. Urban vegetation releases (RE) and absorbs (GEE) the equivalent of 11 and 14%, respectively, of anthropogenic emissions in the most urban portions of the state. While urban vegetation in Massachusetts fully sequesters anthropogenic emissions from smaller cities in the region, Boston's UHI reduces annual C storage by >20% such that vegetation offsets only 2% of anthropogenic emissions. Asynchrony between temporal patterns of biogenic and anthropogenic C fluxes further constrains the emissions mitigation potential of urban vegetation. However, neglecting to account for biogenic C fluxes in cities can impair efforts to accurately monitor, report, verify, and reduce anthropogenic emissions.

A new seasonal-deciduous spring phenology submodel in the Community Land Model 4.5: impacts on carbon and water cycling under future climate scenarios.

A spring phenology model that combines photoperiod with accumulated heating and chilling to predict spring leaf-out dates is optimized using PhenoCam observations and coupled into the Community Land Model (CLM) 4.5. In head-to-head comparison (using satellite data from 2003 to 2013 for validation) for model grid cells over the Northern Hemisphere deciduous broadleaf forests (5.5 million km2 ), we found that the revised model substantially outperformed the standard CLM seasonal-deciduous spring phenology submodel at both coarse (0.9 × 1.25°) and fine (1 km) scales. The revised model also does a better job of representing recent (decadal) phenological trends observed globally by MODIS, as well as long-term trends (1950-2014) in the PEP725 European phenology dataset. Moreover, forward model runs suggested a stronger advancement (up to 11 days) of spring leaf-out by the end of the 21st century for the revised model. Trends toward earlier advancement are predicted for deciduous forests across the whole Northern Hemisphere boreal and temperate deciduous forest region for the revised model, whereas the standard model predicts earlier leaf-out in colder regions, but later leaf-out in warmer regions, and no trend globally. The earlier spring leaf-out predicted by the revised model resulted in enhanced gross primary production (up to 0.6 Pg C yr-1 ) and evapotranspiration (up to 24 mm yr-1 ) when results were integrated across the study region. These results suggest that the standard seasonal-deciduous submodel in CLM should be reconsidered, otherwise substantial errors in predictions of key land-atmosphere interactions and feedbacks may result.

Multiscale modeling of spring phenology across Deciduous Forests in the Eastern United States.

Phenological events, such as bud burst, are strongly linked to ecosystem processes in temperate deciduous forests. However, the exact nature and magnitude of how seasonal and interannual variation in air temperatures influence phenology is poorly understood, and model-based phenology representations fail to capture local- to regional-scale variability arising from differences in species composition. In this paper, we use a combination of surface meteorological data, species composition maps, remote sensing, and ground-based observations to estimate models that better represent how community-level species composition affects the phenological response of deciduous broadleaf forests to climate forcing at spatial scales that are typically used in ecosystem models. Using time series of canopy greenness from repeat digital photography, citizen science data from the USA National Phenology Network, and satellite remote sensing-based observations of phenology, we estimated and tested models that predict the timing of spring leaf emergence across five different deciduous broadleaf forest types in the eastern United States. Specifically, we evaluated two different approaches: (i) using species-specific models in combination with species composition information to 'upscale' model predictions and (ii) using repeat digital photography of forest canopies that observe and integrate the phenological behavior of multiple representative species at each camera site to calibrate a single model for all deciduous broadleaf forests. Our results demonstrate variability in cumulative forcing requirements and photoperiod cues across species and forest types, and show how community composition influences phenological dynamics over large areas. At the same time, the response of different species to spatial and interannual variation in weather is, under the current climate regime, sufficiently similar that the generic deciduous forest model based on repeat digital photography performed comparably to the upscaled species-specific models. More generally, results from this analysis demonstrate how in situ observation networks and remote sensing data can be used to synergistically calibrate and assess regional parameterizations of phenology in models.

Greenness indices from digital cameras predict the timing and seasonal dynamics of canopy-scale photosynthesis.

The proliferation of digital cameras co-located with eddy covariance instrumentation provides new opportunities to better understand the relationship between canopy phenology and the seasonality of canopy photosynthesis. In this paper we analyze the abilities and limitations of canopy color metrics measured by digital repeat photography to track seasonal canopy development and photosynthesis, determine phenological transition dates, and estimate intra-annual and interannual variability in canopy photosynthesis. We used 59 site-years of camera imagery and net ecosystem exchange measurements from 17 towers spanning three plant functional types (deciduous broadleaf forest, evergreen needleleaf forest, and grassland/crops) to derive color indices and estimate gross primary productivity (GPP). GPP was strongly correlated with greenness derived from camera imagery in all three plant functional types. Specifically, the beginning of the photosynthetic period in deciduous broadleaf forest and grassland/crops and the end of the photosynthetic period in grassland/crops were both correlated with changes in greenness; changes in redness were correlated with the end of the photosynthetic period in deciduous broadleaf forest. However, it was not possible to accurately identify the beginning or ending of the photosynthetic period using camera greenness in evergreen needleleaf forest. At deciduous broadleaf sites, anomalies in integrated greenness and total GPP were significantly correlated up to 60 days after the mean onset date for the start of spring. More generally, results from this work demonstrate that digital repeat photography can be used to quantify both the duration of the photosynthetically active period as well as total GPP in deciduous broadleaf forest and grassland/crops, but that new and different approaches are required before comparable results can be achieved in evergreen needleleaf forest.

Direct human influence on atmospheric CO2 seasonality from increased cropland productivity.

Ground- and aircraft-based measurements show that the seasonal amplitude of Northern Hemisphere atmospheric carbon dioxide (CO2) concentrations has increased by as much as 50 per cent over the past 50 years. This increase has been linked to changes in temperate, boreal and arctic ecosystem properties and processes such as enhanced photosynthesis, increased heterotrophic respiration, and expansion of woody vegetation. However, the precise causal mechanisms behind the observed changes in atmospheric CO2 seasonality remain unclear. Here we use production statistics and a carbon accounting model to show that increases in agricultural productivity, which have been largely overlooked in previous investigations, explain as much as a quarter of the observed changes in atmospheric CO2 seasonality. Specifically, Northern Hemisphere extratropical maize, wheat, rice, and soybean production grew by 240 per cent between 1961 and 2008, thereby increasing the amount of net carbon uptake by croplands during the Northern Hemisphere growing season by 0.33 petagrams. Maize alone accounts for two-thirds of this change, owing mostly to agricultural intensification within concentrated production zones in the midwestern United States and northern China. Maize, wheat, rice, and soybeans account for about 68 per cent of extratropical dry biomass production, so it is likely that the total impact of increased agricultural production exceeds the amount quantified here.

Influence of spring and autumn phenological transitions on forest ecosystem productivity.

We use eddy covariance measurements of net ecosystem productivity (NEP) from 21 FLUXNET sites (153 site-years of data) to investigate relationships between phenology and productivity (in terms of both NEP and gross ecosystem photosynthesis, GEP) in temperate and boreal forests. Results are used to evaluate the plausibility of four different conceptual models. Phenological indicators were derived from the eddy covariance time series, and from remote sensing and models. We examine spatial patterns (across sites) and temporal patterns (across years); an important conclusion is that it is likely that neither of these accurately represents how productivity will respond to future phenological shifts resulting from ongoing climate change. In spring and autumn, increased GEP resulting from an 'extra' day tends to be offset by concurrent, but smaller, increases in ecosystem respiration, and thus the effect on NEP is still positive. Spring productivity anomalies appear to have carry-over effects that translate to productivity anomalies in the following autumn, but it is not clear that these result directly from phenological anomalies. Finally, the productivity of evergreen needleleaf forests is less sensitive to phenology than is productivity of deciduous broadleaf forests. This has implications for how climate change may drive shifts in competition within mixed-species stands.

Large seasonal swings in leaf area of Amazon rainforests.

Despite early speculation to the contrary, all tropical forests studied to date display seasonal variations in the presence of new leaves, flowers, and fruits. Past studies were focused on the timing of phenological events and their cues but not on the accompanying changes in leaf area that regulate vegetation-atmosphere exchanges of energy, momentum, and mass. Here we report, from analysis of 5 years of recent satellite data, seasonal swings in green leaf area of approximately 25% in a majority of the Amazon rainforests. This seasonal cycle is timed to the seasonality of solar radiation in a manner that is suggestive of anticipatory and opportunistic patterns of net leaf flushing during the early to mid part of the light-rich dry season and net leaf abscission during the cloudy wet season. These seasonal swings in leaf area may be critical to initiation of the transition from dry to wet season, seasonal carbon balance between photosynthetic gains and respiratory losses, and litterfall nutrient cycling in moist tropical forests.