Leveraging Earth Engine for Transforming Land Cover & Above Ground Biomass into Greenhouse Gas Emissions Ensembles
Explore the source record for details and available documents.
Earth observation↗
SEARCH · Search NASA
Results for “forest biomass”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Synthesizing Disparate LiDAR and Satellite Datasets through Deep Learning to Generate Wall-to-Wall Regional Inventories for the Complex, Mixed-Species Forests of the Eastern United States
Light detection and ranging (LiDAR) has become a commonly-used tool for generating remotely-sensed forest inventories. However, LiDAR-derived forest inventories have remained uncommon at a regional scale due to varying parameters among LiDAR data acquisitions and the availability of sufficient calibration data. Here, we present a model using a 3-D convolutional neural network (CNN), a form of deep learning capable of scanning a LiDAR point cloud, combined with coincident satellite data (spectral, phenology, and disturbance history). We compared this approach to traditional modeling used for making forest predictions from LiDAR data (height metrics and random forest) and found that the CNN had consistently lower uncertainty. We then applied the CNN to public data over six New England states in the USA, generating maps of 14 forest attributes at a 10 m resolution over 85% of the region. Aboveground biomass estimates produced a root mean square error of 36 Mg ha−1 (44%) and were within the 97.5% confidence of independent county-level estimates for 33 of 38 or 86.8% of the counties examined. CNN predictions for stem density and percentage of conifer attributes were moderately successful, while predictions for detailed species groupings were less successful. The approach shows promise for improving the prediction of forest attributes from regional LiDAR data and for combining disparate LiDAR datasets into a common framework for large-scale estimation.
Elias Ayrey↗
Fusion of AIRSAR and TM Data for Parameter Classification and Estimation in Dense and Hilly Forests
The expanded remotely sensed data space consisting of coincident radar backscatter and optical reflectance data provides for a more complete description of the Earth surface. This is especially useful where many parameters are needed to describe a certain scene, such as in the presence of dense and complex-structured vegetation or where there is considerable underlying topography. The goal of this paper is to use a combination of radar and optical data to develop a methodology for parameter classification for dense and hilly forests, and further, class-specific parameter estimation. The area to be used in this study is the H. J. Andrews Forest in Oregon, one of the Long-Term Ecological Research (LTER) sites in the US. This area consists of various dense old-growth conifer stands, and contains significant topographic relief. The Andrews forest has been the subject of many ecological studies over several decades, resulting in an abundance of ground measurements. Recently, biomass and leaf-area index (LAI) values for approximately 30 reference stands have also become available which span a large range of those parameters. The remote sensing data types to be used are the C-, L-, and P-band polarimetric radar data from the JPL airborne SAR (AIRSAR), the C-band single-polarization data from the JPL topographic SAR (TOPSAR), and the Thematic Mapper (TM) data from Landsat, all acquired in late April 1998. The total number of useful independent data channels from the AIRSAR is 15 (three frequencies, each with three unique polarizations and amplitude and phase of the like-polarized correlation), from the TOPSAR is 2 (amplitude and phase of the interferometric correlation), and from the TM is 6 (the thermal band is not used). The range pixel spacing of the AIRSAR is 3.3m for C- and L-bands and 6.6m for P-band. The TOPSAR pixel spacing is 10m, and the TM pixel size is 30m. To achieve parameter classification, first a number of parameters are defined which are of interest to ecologists for forest process modeling. These parameters include total biomass, leaf biomass, LAI, and tree height. The remote sensing data from radar and TM are used to formulate a multivariate analysis problem given the ground measurements of the parameters. Each class of each parameter is defined by a probability density function (pdf), the spread of which defines the range of that class. High classification accuracy results from situations in which little overlap occurs between pdfs. Classification results provide the basis for the future work of class-specific parameter estimation using radar and optical data. This work was performed in part by the Jet Propulsion Laboratory, California Institute of Technology, Pasadena, CA, and in part by the NASA Ames Research Center, Moffett Field, CA, both under contract from the National Aeronautics and Space Administration.
Moghaddam, Mahta↗
Habitability of the Earth
Test methods to measure land vegetation biomass, net biological production and leaf area index by remote sensing are applied to estimate the biomass and net biological productivity of selected biomes, including the boreal forests and north temperate grasslands. Field verification is conducted in conjunction with remote sensing of pertinent variables and the development of a data base of relevant material initiated. Measurements are made in the Superior National Forest, Minnesota with the following remote sensing instruments: (1) an airborne 8 band Barnes radiometer; (2) the thematic mapper simulator flown in a NASA C-130 aircraft; (3) MSS data from LANDSAT 3; and (4) AVHRR data from the NOAA Satellite. Vegetation data collected as field verification in the Superior National Forest is summarized.
Botkin, D. B.↗
Studies of optical and biological properties of terrestrial land cover using multispectral linear array technology
A series of experiments to study the optical and biological properties of terrestrial land cover are planned for late 1987 using a six-channel imaging spectroradiometer based on newly developed multispectral linear array (MLA) detector technology. Data from selected portions of the Sahel and rain forests of Africa and South America will be used to delineate biomass classes and estimate spherical albedos. A spatial resolution of 15 meters in the four visible-near IR channels and 30 meters in two shortwave IR channels, including a 'new' channel centered at 1.24 micrometers when combined with a spectral width of 20 nm for all channels, will be used to investigate possible improvements in land cover classification. Technology demonstrations include a test of data compression on data quality, the first spaceborne utilization of short wave infrared Schottky barrier Pd2Si detector arrays, and the use of close-butted, multi-array modules with attached spectral filters.
Barnes, W. L.↗
Disturbance and the Carbon Balance of US Forests: A Quantitative Review of Impacts from Harvests, Fires, Insects, and Droughts
Disturbances are a major determinant of forest carbon stocks and uptake. They generally reduce land carbon stocks but also initiate a regrowth legacy that contributes substantially to the contemporary rate of carbon stock increase in US forestlands. As managers and policy makers increasingly look to forests for climate protection and mitigation, and because of increasing concern about changes in disturbance intensity and frequency, there is a need for synthesis and integration of current understanding about the role of disturbances and other processes in governing forest carbon cycle dynamics, and the likely future of this and other sinks for atmospheric carbon. This paper aims to address that need by providing a quantitative review of the distribution, extent and carbon impacts of the major disturbances active in the US. We also review recent trends in disturbances, climate, and other global environmental changes and consider their individual and collective contributions to the US carbon budget now and in the likely future. Lastly, we identify some key challenges and opportunities for future research needed to improve current understanding, advance predictive capabilities, and inform forest management in the face of these pressures. Harvest is found to be the most extensive disturbance both in terms of area and carbon impacts, followed by fire, windthrow and bark beetles, and lastly droughts. Collectively these lead to the gross loss of about 200 Tg C y(exp -1) in live biomass annually across the conterminous US. At the same time, the net change in forest carbon stocks is positive (190 Tg C y(exp -1)), indicating not only forest resilience but also an apparently large response to growth enhancements such as fertilization by CO2 and nitrogen. Uncertainty about disturbance legacies, disturbance interactions, likely trends, and global change factors make the future of the US forest carbon sink unclear. While there is scope for management to enhance carbon sinks in US forests, tradeoffs with other values and uses are likely to significantly limit practical implementation. Continued and expanded remote sensing and field-based monitoring capabilities and manipulative experimentation are needed to improve understanding of the US forest carbon sink, and assess how disturbance processes are responding to the pressures of global environmental change. In addition, continued development and application of holistic, decision support tools that consider a range of forest values are needed to enable managers and policy makers to use the best available information for guiding forest resources now and into the future.
Williams, Christopher A.↗
Optical Properties of Boreal Region Biomass Burning Aerosols in Central Alaska and Seasonal Variation of Aerosol Optical Depth at an Arctic Coastal Site
Long-term monitoring of aerosol optical properties at a boreal forest AERONET site in interior Alaska was performed from 1994 through 2008 (excluding winter). Large interannual variability was observed, with some years showing near background aerosol optical depth (AOD) levels (<0.1 at 500 nm) while 2004 and 2005 had August monthly means similar in magnitude to peak months at major tropical biomass burning regions. Single scattering albedo (omega (sub 0); 440 nm) at the boreal forest site ranged from approximately 0.91 to 0.99 with an average of approximately 0.96 for observations in 2004 and 2005. This suggests a significant amount of smoldering combustion of woody fuels and peat/soil layers that would result in relatively low black carbon mass fractions for smoke particles. The fine mode particle volume median radius during the heavy burning years was quite large, averaging approximately 0.17 micron at AOD(440 nm) = 0.1 and increasing to approximately 0.25 micron at AOD(440 nm) = 3.0. This large particle size for biomass burning aerosols results in a greater relative scattering component of extinction and, therefore, also contributes to higher omega (sub 0). Additionally, monitoring at an Arctic Ocean coastal site (Barrow, Alaska) suggested transport of smoke to the Arctic in summer resulting in individual events with much higher AOD than that occurring during typical spring Arctic haze. However, the springtime mean AOD(500 nm) is higher during late March through late May (approximately 0.150) than during summer months (approximately 0.085) at Barrow partly due to very few days with low background AOD levels in spring compared with many days with clean background conditions in summer.
Eck, T. F.↗
Trace gas and particulate emissions from biomass burning in temperate ecosystems
Emissions measured from fires in graminoid wetlands, Mediterranean chaparrals, and boreal forests, suggest that such ecosystemic parameters as fuel size influence combustion emissions in ways that are broadly predictable. The degree of predictability is most noticeable when wetland fire-related results are compared with boreal forest emissions; the inorganic fraction of the particulate emissions is close in composition irrespective of the ecosystem. It is found that both aerosol and trace gas emissions are influenced by the phase of combustion.
Cofer, Wesley R., III↗
Abrupt Increases in Amazonian Tree Mortality Due to Drought-Fire Interactions
Interactions between climate and land-use change may drive widespread degradation of Amazonian forests. High-intensity fires associated with extreme weather events could accelerate this degradation by abruptly increasing tree mortality, but this process remains poorly understood. Here we present, to our knowledge, the first field-based evidence of a tipping point in Amazon forests due to altered fire regimes. Based on results of a large-scale, longterm experiment with annual and triennial burn regimes (B1yr and B3yr, respectively) in the Amazon, we found abrupt increases in fire-induced tree mortality (226 and 462%) during a severe drought event, when fuel loads and air temperatures were substantially higher and relative humidity was lower than long-term averages. This threshold mortality response had a cascading effect, causing sharp declines in canopy cover (23 and 31%) and aboveground live biomass (12 and 30%) and favoring widespread invasion by flammable grasses across the forest edge area (80 and 63%), where fires were most intense (e.g., 220 and 820 kW x m(exp −1)). During the droughts of 2007 and 2010, regional forest fires burned 12 and 5% of southeastern Amazon forests, respectively, compared with less than 1% in nondrought years. These results show that a few extreme drought events, coupled with forest fragmentation and anthropogenic ignition sources, are already causing widespread fire-induced tree mortality and forest degradation across southeastern Amazon forests. Future projections of vegetation responses to climate change across drier portions of the Amazon require more than simulation of global climate forcing alone and must also include interactions of extreme weather events, fire, and land-use change.
MODIS↗
Emissions of Black Carbon, Organic, and Inorganic Aerosols From Biomass Burning in North America and Asia in 2008
Reliable assessment of the impact of aerosols emitted from boreal forest fires on the Arctic climate necessitates improved understanding of emissions and the microphysical properties of carbonaceous (black carbon (BC) and organic aerosols (OA)) and inorganic aerosols. The size distributions of BC were measured by an SP2 based on the laser-induced incandescence technique on board the DC-8 aircraft during the NASA ARCTAS campaign. Aircraft sampling was made in fresh plumes strongly impacted by wildfires in North America (Canada and California) in summer 2008 and in those transported from Asia (Siberia in Russia and Kazakhstan) in spring 2008. We extracted biomass burning plumes using particle and tracer (CO, CH3CN, and CH2Cl2) data. OA constituted the dominant fraction of aerosols mass in the submicron range. The large majority of the emitted particles did not contain BC. We related the combustion phase of the fire as represented by the modified combustion efficiency (MCE) to the emission ratios between BC and other species. In particular, we derived the average emission ratios of BC/CO = 2.3 +/- 2.2 and 8.5 +/- 5.4 ng/cu m/ppbv for BB in North America and Asia, respectively. The difference in the BC/CO emission ratios is likely due to the difference in MCE. The count median diameters and geometric standard deviations of the lognormal size distribution of BC in the BB plumes were 136-141 nm and 1.32-1.36, respectively, and depended little on MCE. These BC particles were thickly coated, with shell/core ratios of 1.3-1.6. These parameters can be used directly for improving model estimates of the impact of BB in the Arctic.
Kondo, Y.↗
Central Africa Energy: Utilizing NASA Earth Observations to Explore Flared Gas as an Energy Source Alternative to Biomass in Central Africa
Much of Central Africa's economy is centered on oil production. Oil deposits lie below vast amounts of compressed natural gas. The latter is often flared off during oil extraction due to a lack of the infrastructure needed to utilize it for productive energy generation. Though gas flaring is discouraged by many due to its contributions to greenhouse emissions, it represents a waste process and is rarely tracked or recorded in this region. In contrast to this energy waste, roughly 80% of Africa's population lacks access to electricity and in turn uses biomass such as wood for heat and light. In addition to the dangers incurred from collecting and using biomass, the practice commonly leads to ecological change through the acquisition of wood from forests surrounding urban areas. The objective of this project was to gain insight on domestic energy usage in Central Africa, specifically Angola, Gabon, and the Republic of Congo. This was done through an analysis of deforestation, an estimation of gas flared, and a suitability study for the infrastructure needed to realize the natural gas resources. The energy from potential natural gas production was compared to the energy equivalent of the biomass being harvested. A site suitability study for natural gas pipeline routes from flare sites to populous locations was conducted to assess the feasibility of utilizing natural gas for domestic energy needs. Analyses and results were shared with project partners, as well as this project's open source approach to assessing the energy sector. Ultimately, Africa's growth demands energy for its people, and natural gas is already being produced by the flourishing petroleum industry in numerous African countries. By utilizing this gas, Africa could reduce flaring, recuperate the financial and environmental loss that flaring accounts for, and unlock a plentiful domestic energy source for its people. II. Introduction Background Africa is home to numerous burgeoning economies; a significant number rely on oil production as their primary source of revenue. Relative to its size and population density, the continent has a wealth of natural resources, including oil and natural gas deposits. The exploration of these resources is not a new endeavor, but rather one that spans decades, up to a century in some places. Their resources, if realized, could provide a great means of economic and social mobility for the people of Africa. Currently, Africa represents about 12 % of the energy market, yet at the same time, consumes only 3 % of the world's energy (Kasekende 2009). The higher
Jones, Amber↗
A Multi-scaled Analysis of Forest Structure using Individual-Based Modeling in a Costa Rican Rainforest
Consideration of scale is essential when examining structural relationships in forests. In this study, we present a parameterization of the FORMIND individual-based forest model for old growth Atlantic lowland rainforest in La Selva, Costa Rica. Results show that the simulated forest reproduces the structural complexity of Costa Rican rainforest within 2.3% of aboveground biomass values, based on comparisons with CARBONO inventory plot data. The Costa Rica FORMIND simulation was then used to investigate the relationship between canopy height and aboveground biomass (AGB), leaf area index (LAI) and gross primary productivity (GPP) at different spatial scales (20 × 20 m, 60 × 60 m, 100mx100m). The relationship between aboveground biomass and height is of particular importance toward the calibration of various remote sensing products including lidar and radar, whereas the LAI and GPP relationships are understudied in this context. We found that the relationship between all three variables and height varies considerably: the relationship is stronger at finer scales and weaker at coarser resolution. However, in all three comparisons, RMSE also decreased as scales coarsened, with the largest difference shown between 100 m and 10 m resolutions in relating AGB to Lorey's height (R^2 decreased by 0.3; RMSE decreased by 114.5 Mg/ha). This suggests that a trade-off between accuracy and precision exists, and further highlights the importance of spatial scale in determining the relatability of forest structure variables.
A.H.Armstrong↗
Mapping Forest Disturbance Intensity in North and South Carolina Using Annual Landsat Observations and Field Inventory Data
Disturbance and regrowth are vital processes in determining the roles of forest ecosystem in the carbon and biogeochemical cycles. Using time series observations, the vegetation change tracker (VCT) algorithm was designed to map the location, timing, and spectral magnitudes of forest disturbance events. While these spectral disturbance magnitudes are indicative of physical changes in tree cover or biomass, their quantitative relationships have yet to be established. This study focuses on estimating disturbance intensity as measured by percent basal area removal using spectral indices from the VCT algorithm over North and South Carolina. Repeat measurements on Forest Service Forest Inventory Analysis (FIA) ground plots, which provide changes in basal area between multiple dates at precise locations, are used for training and validation of the model. The overall R2 between predicted disturbance intensity and reference data is 0.66, and cross-validation prediction uncertainty is 14% in North Carolina. Possible causes of this uncertainty could be site heterogeneity and the temporal offset between ground measurements and satellite observations. Results show the area of stand clearing disturbances remains relatively stable around 1143 km2 yr−1 in North and South Carolina throughout the period of observations (1985–2015). The average amount of forest area affected by partial disturbance is much higher at 3287 km2 yr−1. The area of partial disturbances has strong inter-annual variability with a high value of 6000 km2 in 2007 and a low value of 1919 km2 in 2013.
Xin Tao↗
Estimating Volume, Biomass, and Carbon in Hedmark County, Norway Using a Profiling LiDAR
A profiling airborne LiDAR is used to estimate the forest resources of Hedmark County, Norway, a 27390 square kilometer area in southeastern Norway on the Swedish border. One hundred five profiling flight lines totaling 9166 km were flown over the entire county; east-west. The lines, spaced 3 km apart north-south, duplicate the systematic pattern of the Norwegian Forest Inventory (NFI) ground plot arrangement, enabling the profiler to transit 1290 circular, 250 square meter fixed-area NFI ground plots while collecting the systematic LiDAR sample. Seven hundred sixty-three plots of the 1290 plots were overflown within 17.8 m of plot center. Laser measurements of canopy height and crown density are extracted along fixed-length, 17.8 m segments closest to the center of the ground plot and related to basal area, timber volume and above- and belowground dry biomass. Linear, nonstratified equations that estimate ground-measured total aboveground dry biomass report an R(sup 2) = 0.63, with an regression RMSE = 35.2 t/ha. Nonstratified model results for the other biomass components, volume, and basal area are similar, with R(sup 2) values for all models ranging from 0.58 (belowground biomass, RMSE = 8.6 t/ha) to 0.63. Consistently, the most useful single profiling LiDAR variable is quadratic mean canopy height, h (sup bar)(sub qa). Two-variable models typically include h (sup bar)(sub qa) or mean canopy height, h(sup bar)(sub a), with a canopy density or a canopy height standard deviation measure. Stratification by productivity class did not improve the nonstratified models, nor did stratification by pine/spruce/hardwood. County-wide profiling LiDAR estimates are reported, by land cover type, and compared to NFI estimates.
Nelson, Ross↗
Forest Understory Fire in the Brazilian Amazon in ENSO and Non-ENSO Years: Area Burned and Committed Carbon Emissions
"Understory fires" that burn the floor of standing forests are one of the most important types of forest impoverishment in the Amazon, especially during the severe droughts of El Nino Southern Oscillation (ENSO) episodes. However, we are aware of no estimates of the areal extent of these fires for the Brazilian Amazon and, hence, of their contribution to Amazon carbon fluxes to the atmosphere. We calculated the area of forest understory fires for the Brazilian Amazon region during an El Nino (1998) and a non El Nino (1995) year based on forest fire scars mapped with satellite images for three locations in eastern and southern Amazon, where deforestation is concentrated. The three study sites represented a gradient of both forest types and dry season severity. The burning scar maps were used to determine how the percentage of forest that burned varied with distance from agricultural clearings. These spatial functions were then applied to similar forest/climate combinations outside of the study sites to derive an initial estimate for the Brazilian Amazon. Ninety-one percent of the forest area that burned in the study sites was within the first kilometer of a clearing for the non ENSO year and within the first four kilometers for the ENSO year. The area of forest burned by understory forest fire during the severe drought (ENSO) year (3.9 millions of hectares) was 13 times greater than the area burned during the average rainfall year (0.2 million hectares), and twice the area of annual deforestation rate. Dense forest was, proportionally, the forest area most affected by understory fires during the El Nino year, while understory fires were concentrated in transitional forests during the year of average rainfall. Our estimate of aboveground tree biomass killed by fire ranged from 0.06 Pg to 0.38 Pg during the ENSO and from 0,004 Pg to 0,024 Pg during the non ENSO.
Alencar, A.↗
Vegetation Heights for Landsat Land Cover Classes in the Interior Boreal Forest Region of Alaska
Accurate determination and mapping of vegetation height is a prerequisite for regional estimations of above ground biomass, carbon pool dynamics, and improved modeling of future climate-vegetation-soil interactions. This study combined land cover class maps (at 30-m pixel resolution) derived from Landsat image bands with canopy vegetation heights detected at 1-m resolution from NASA G-LiHT Lidar flights, both from the year 2014 for the Middle Tanana Valley near Fairbanks, Alaska. The statistical distributions and structural attributes from Lidar plant heights were derived and refined for all Landsat vegetation classes. Results showed that the tallest cover classes in this region were deciduous forests (dominated by birch and aspen), followed by evergreen spruce and wooded wetlands. Open shrubland and herbaceous cover areas were determined to be 2 - 5 m shorter, on average, than these boreal forest stands. Tall shrubland cover areas were slightly shorter than most spruce-dominated evergreen forest areas. Nearly 60% of all low shrubland and herbaceous wetland pixels were detected with at least one tree crown greater than 10 m tall. This was in contrast to evergreen and deciduous forest classes that showed 88% of stands with at least one tree crown greater than 10 m tall.
Vegetation↗
Aboveground carbon loss associated with the spread of ghost forests as sea levels rise
Coastal forests sequester and store more carbon than their terrestrial counterparts but are at greater risk of conversion due to sea level rise. Saltwater intrusion from sea level rise converts freshwater-dependent coastal forests to more salt-tolerant marshes, leaving 'ghost forests' of standing dead trees behind. Although recent research has investigated the drivers and rates of coastal forest decline, the associated changes in carbon storage across large extents have not been quantified. We mapped ghost forest spread across coastal North Carolina, USA, using repeat Light Detection and Ranging (LiDAR) surveys, multi-temporal satellite imagery, and field measurements of aboveground biomass to quantify changes in aboveground carbon. Between 2001 and 2014, 15% (167 km2) of unmanaged public land in the region changed from coastal forest to transition-ghost forest characterized by salt-tolerant shrubs and herbaceous plants. Salinity and proximity to the estuarine shoreline were significant drivers of these changes. This conversion resulted in a net aboveground carbon decline of 0.13 ± 0.01 TgC. Because saltwater intrusion precedes inundation and influences vegetation condition in advance of mature tree mortality, we suggest that aboveground carbon declines can be used to detect the leading edge of sea level rise. Aboveground carbon declines along the shoreline were offset by inland aboveground carbon gains associated with natural succession and forestry activities like planting (2.46 ± 0.25 TgC net aboveground carbon across study area). Our study highlights the combined effects of saltwater intrusion and land use on aboveground carbon dynamics of temperate coastal forests in North America. By quantifying the effects of multiple interacting disturbances, our measurement and mapping methods should be applicable to other coastal landscapes experiencing saltwater intrusion. As sea level rise increases the landward extent of inundation and saltwater exposure, investigations at these large scales are requisite for effective resource allocation for climate adaptation. In this changing environment, human intervention, whether through land preservation, restoration, or reforestation, may be necessary to prevent aboveground carbon loss.
ghost forests↗
Satellite-Based Evidence for Shrub and Graminoid Tundra Expansion in Northern Quebec from 1986-2010
Global vegetation models predict rapid poleward migration of tundra and boreal forest vegetation in response to climate warming. Local plot and air-photo studies have documented recent changes in high-latitude vegetation composition and structure, consistent with warming trends. To bridge these two scales of inference, we analyzed a 24-year (1986-2010) Landsat time series in a latitudinal transect across the boreal forest-tundra biome boundary in northern Quebec province, Canada. This region has experienced rapid warming during both winter and summer months during the last forty years. Using a per-pixel (30 m) trend analysis, 30% of the observable (cloud-free) land area experienced a significant (p < 0.05) positive trend in the Normalized Difference Vegetation Index (NDVI). However, greening trends were not evenly split among cover types. Low shrub and graminoid tundra contributed preferentially to the greening trend, while forested areas were less likely to show significant trends in NDVI. These trends reflect increasing leaf area, rather than an increase in growing season length, because Landsat data were restricted to peak-summer conditions. The average NDVI trend (0.007/yr) corresponds to a leaf-area index (LAI) increase of ~0.6 based on the regional relationship between LAI and NDVI from the Moderate Resolution Spectroradiometer (MODIS). Across the entire transect, the area-averaged LAI increase was ~0.2 during 1986-2010. A higher area-averaged LAI change (~0.3) within the shrub-tundra portion of the transect represents a 20-60% relative increase in LAI during the last two decades. Our Landsat-based analysis subdivides the overall high-latitude greening trend into changes in peak-summer greenness by cover type. Different responses within and among shrub, graminoid, and tree-dominated cover types in this study indicate important fine-scale heterogeneity in vegetation growth. Although our findings are consistent with community shifts in low-biomass vegetation types over multi-decadal time scales, the response in tundra and forest ecosystems to recent warming was not uniform.
McManus, K. M.↗