Terrestrial Carbon Cycle

My research contributes to NASA’s mission: ‘to advance and communicate scientific knowledge and understanding of the earth, the solar system, and the universe’. Two NASA earth science questions are at the center of my work: ‘How does the earth system respond to natural and human-induced changes?’ and ‘What are the consequences of change in the earth system for human civilization?’ To explore these questions, I conduct research in three areas: 1) data fusion to develop long term data records of vegetation dynamics for carbon cycle studies; 2) terrestrial ecosystem modeling of land-cover/land-use change in the context of impact to carbon storage; and 3) the development of models and methods that enable the quantification of the impact of climate change on ecosystem carbon storage. During the next five years, I plan to study and produce publications on the carbon consequences of land-cover/land-use change associated to agriculture production practices, fire, insect herbivory outbreaks, drought, climate warming impacting soil processes and plant productivity in a number of regions around the globe.

Disturbance Mapping

· Investigate long-term (greater than 25 years) global remote sensing observations and simulate net biome production (calculation of the total carbon balance including human initiated disturbances) at fine-scales.

· Combine multi-scale remote sensing observations, so that hotspots can be identified and processes understood to refine our uncertainties associated to the global carbon balance.

Biogeochemistry Modeling

· Simulate ecosystem disturbance phenomena with biogeochemistry modeling to quantitatively understand dynamics is an emerging field with the advent of relatively inexpensive high powered computing and large quantities of remote sensing data.

· Understand spatio-temporal biosphere dynamics in ecosystem carbon balance assessments that were previously overlooked due to underlying inadequacies of available course resolution datasets and limited computing power.

Disturbance, Growth, and Recovery of Boreal Forests Spanning the Satellite Era: 3D Structure, Site Index, and Ecosystem Carbon Flux with Changing Climate

Vegetation & Soil

Climate change is altering vegetation productivity, dynamics, and C sequestration at continental scales across the higher northern latitudes. These shifts are reflected in changing vegetation canopy structure (cover and height), which varies from continuous forests of tall trees to discrete, isolated forest patches and evenly distributed expanses of growth-stunted trees. Although variation occurs at spatial scales not resolved by most earth-observing satellites due to local environmental factors, including terrain, winter snow and permafrost active layer depth, wind, and soil characteristics. Recent studies have found that Arctic greening is associated with densification of shrubs, increasing biomass, and establishment of new shrubs within thawing patterned ground features. However, they have not been sufficiently comprehensive to establish a map of growth potential for northern forests.

Understanding site-scale forest structure and its relation to environmental factors can improve broad-scale estimates of boreal C-flux. Previous studies have found large increases in productivity, but turnover rates in models remain a large source of uncertainty. Represented as in forestry as SI, site-specific estimates of forest growth potential will reduce this uncertainty in live C turnover into soil C pools. We propose to expand the NASA Global Forest Cover and Change (GFCC) dataset, correlate it with spatially explicit environmental data layers, and incorporate the relationships into DGVMs to understand environmental constraints on canopy structure and to predict impacts of environmental change on vegetation cover and C stock & flux.

With prior support from the NASA Carbon Cycle Science (CCS) program (13-CARBON13_2-0377 A high-resolution circumpolar delineation of the forest-tundra ecotone with implications for carbon balance), we have developed samples of canopy-cover and -height estimates from airborne LiDAR measurements and photogrammetric stereo image pairs. We have used these high-resolution reference datasets to calibrate and validate the NASA Earth Science Data Record (ESDR) of Global Forest Change (GFCC). These linear calibrations improved the accuracy of the GFCC global 30-m, annual resolution estimates of tree cover across the boreal biome and refined previous models and estimates of the location of the Taiga Tundra Ecotone (TTE). However, there is no pan-boreal, Landsat-resolution record of forest change from 1972 to 2000—a spatial and temporal scale needed to study the dynamics of slow-growing boreal forests.

The proposed study will develop a sample of reference data of > 10,000 sub-meter resolution stereo image pairs and use it to retrieve biome-wide estimates of forest stand age from the entire Landsat archive, from the 1970s until present. This extension of the disturbance record will enable modeling and analysis of forest growth and disturbance across a ~40-year chronosequence—the time span necessary for studying recent greening and browning trends. We will then pair the estimates of stand age, structure, and environmental factors into a chronosequence of SI across northern forests in order to understand local topographic and climatic effects on canopy structure and growth. Widely applied in forest ecology and silviculture, SI is an empirical estimate of forest productivity that reflects the effect of the local environment on vertical growth.

Our objectives are to:
1. Produce a stand-history/forest-age map: using a time-series of forest cover at annual, sub-hectare resolution spanning the entire Landsat archive pan-boreal, from 1972 to present;
2. Model and estimate forest growth potential: using a chronosequence of tree-canopy cover and height from samples of WorldView-1, 2, and 3 stereo imagery to estimate height-growth potential as the forest Site Index (SI); and
3. Estimate pan-boreal forest net C-flux: by incorporating stand age and SI into the Lund-Potsdam-Jena (LPJ) DGVM.

This study leverages our existing team (NASA GSFC and UMD GLCF) with three important resources: (1) access to sub-meter commercial satellite data, 2) NASA Ames stereo pipeline software, and 3) NASA GSFC & UMD GLCF supercomputing. All three aspects provide the means to estimate forest structure at sub-meter resolution and to map long-term, multi-decadal forest growth and productivity across the biome at sub-hectare resolution.

This proposal addresses NASA’s interest in C dynamics of Arctic-Boreal Terrestrial Ecosystems. Our findings will be significant sentinels of climate change and for developing confidence in predictions of changes in C balance due to a rapidly changing environment. Our work will contribute directly to NASA’s interests in characterizing critical ecosystems, especially in high latitudes, and can contribute data and science analysis to the proposed Arctic-Boreal Vulnerability Experiment (ABoVE). DGVMs that require SI and forest age structure can use these results to reduce currently large uncertainties in terrestrial C-flux estimates.

Automated protocols for generating very high-resolution commercial validation products with NASA HEC resources

Remote Sensing

The volume of sub-meter remotely sensed data is growing at rates exceeding petabytes per year. Over the past decade, costs for data storage systems and computing have both dropped exponentially. This has opened the door for “Big Data” processing using spaceborne optical imagery to characterize land surface phenomena in High-End Computing (HEC; acronyms list on page 21) environments. Recent examples include Google Earth Engine, Amazon Web Services ‒ Cloud Computing Services, NASA Earth Exchange and NASA Center for Climate Simulation (NCCS) Advanced Data Analytics Platform (ADAPT). At the same time, a growing constellation of commercial very-high-resolution (VHR) satellites offer global ~1-2-day repeat coverage that can complement NASA Earth Observing (EO) missions with stereo and super-spectral capabilities. Through no-direct-cost licensing agreements the National Geospatial-Intelligence Agency (NGA) and NASA Goddard Spaceflight Center (GSFC) are acquiring petabytes of archived DigitalGlobe (DG) 0.3 ‒ 0.5 m panchromatic and 2 ‒ 4 m multi-spectral imagery from around the globe. The DG archive includes data from 7 satellites including WorldView-1, 2, 3, 4, Quickbird-2, GeoEye-1 and IKONOS-2. Prior to 2008, a limited number of EO VHR images were often used individually for evaluation and validation over small areas, to compare to coarser resolution NASA EO data because these data were snapshots in space and time. Now, the DG constellation enables repeat, contiguous coverage over large regions. These data are a valuable EO resource for scaling global ecological and geological phenomena occurring at sub-meter scales that can enhance NASA EO science that occurs at moderate to coarse resolutions. This proposal seeks to provide tools as an Application Program Interface (API) for mass processing spatially contiguous and temporally consistent archived NASA-GSFC DG VHR data that can only efficiently be performed on NASA HEC resources due to DG-NGA licensing limitations and computational requirements.

We will develop an API for generating VHR products to support NASA funded scientists and NASA EO missions that include two primary foci:

1) On Demand VHR Regional Mosaics ‒ Systematic ortho-rectified and co-registered multi-temporal, panchromatic 0.3 ‒ 0.5 m and unsharpened 2-m multi-spectral imagery compiled as user defined regional mosaics will allow for spatially continuous and temporally consistent reference. Such reference will provide an easily accessible calibration and evaluation dataset for NASA funded scientists. This work builds upon PI Neigh (http://cad4nasa.gsfc.nasa.gov) and Co-I’s Carroll, Slayback, Montesano, and Tucker’s experience in processing VHR data on GSFC’s NCCS ADAPT cluster. We will work with Co-I Lyapustin to generate surface reflectance data from VHR imagery over pseudo-invariant calibration sites for cross-calibration with Landsat and MODIS to minimize the effects of topography, view angle, date and time of day of collection. We will also develop a process for mosaicing and normalizing ortho-rectified images to create scientific data products useful for many NASA programmatic activities, including biodiversity, tree canopy closure, surface water fraction, and cropped area for smallholder agriculture among others.

2) On Demand VHR DEMs – Systematic processing of available along- and cross-track stereo VHR imagery to produce VHR Digital Elevation Models (DEMs) using the NASA Ames Stereo Pipeline open source software (https://ti.arc.nasa.gov/tech/asr/intelligent-robotics/ngt/stereo/). Co-I’s Shean, Alexandrov and Montesano will work together to apply a systematic DEM co-registration approach to generate products with < 0.5 ‒ 1.0 m horizontal and vertical accuracy that can support NASA missions and a number of different science programs. These include Earth surface studies within the Cryosphere (e.g., glacier mass balance, ice flow rates, snow depth), Hydrosphere (e.g., lake/water body levels, surface elevation dynamics, e.g., subsidence from groundwater depletion, thermokarst etc.), Biosphere (e.g., land-cover land-use-change, forest structure, canopy height/cover), and natural hazards (e.g., volcanoes, landslides, earthquakes among others).

NASA Earth Science has underutilized DG VHR imagery due to the “Big Data” infrastructure necessary but generally unavailable to individual NASA PIs. Successful development of a NASA HEC protocol to process VHR data could surmount the current VHR requirements of mass storage and parallel processing. DG VHR imagery requires baseline software and HEC to produce VHR products that will have broad benefits to NASA Earth Science programs.
 

High Mountain Asia and Beyond: Regional Changes in Climate, Glaciers and Water Resources

Hydrology / Water Cycle

Glacier mass loss in High Mountain Asia (HMA) may have far-reaching consequences for the region’s water resources. These changes in the HMA cryosphere are critical for the well-being of the people relying on the these freshwater resources in the vast downstream regions beyond the HMA itself. It is vital that such research be carried out now in order to allow us to estimate these transformations and to better arm the regional stakeholders with improved information and tools thus enabling improved planning and adaptation strategies. However the physical constraints, processes and dependencies of this unique climate-glacier-hydrology system are not well understood and large uncertainties of the glacio-hydrological response to future climate change remain due to the lack of detailed meteorological, glacier and hydrological data, the heterogeneity of climate processes, inaccessible terrain, and the topographical extreme relief. The proposed work aims to provide an integrated framework for the entire HMA region, suitable for understanding past changes in glacier mass and associated streamflow in response to climate change and for projecting those changes into the future combining extensive modeling with relevant glacier products from remote sensing.

Specifically we will a) use visible and radar remote sensing products to derive glacier volume changes, snow line altitudes and debris; b) apply a regional climate model with unprecedented spatial resolution to elucidate the regional-scale monsoon-driven climate dynamics with focus on precipitation patterns across the HMA region, c) model recent glacier changes and forecast future glacier evolution, and d) quantify the hydrological response to climate and glacier changes and forecast how those changes impact human water availability downstream of HMA. In addition we will develop and provide a rich set of integrated operational tools for assessing and forecasting regional-scale climate-driven glacier and hydrological changes for direct integration into the Glacial Melt Toolbox (GMELT) coordinated by NASA’s High Mountain Asia Team (HiMAT).

This project will, for the first time, integrate high-resolution modeling of the climate heterogeneity in HMA with regional-scale glacier hydrological modeling specifically adjusted to HMA and informed by a suite of observations from in-situ and satellite-derived data. High resolution stereo imagery from DigitalGlobe satellites will generate glacier volume change calculations, while Landsat, MODIS, ASTER, VIIRS, ALOS-PALSAR1/2 and Sentinel1/1b satellites will be used for snow cover, and glacier debris cover mapping over the glaciers, and GRACE data will inform glacier mass balance modeling. The project will foster collaboration between an interdisciplinary team of scientists from NASA, the University of Alaska and the University of New Hampshire with extensive experience in visible and radar remote sensing, high-resolution atmospheric modeling, glacier modeling, and regional-scale hydrological modeling.
 

Commercial Hi-res Data Distribution

Remote Sensing

The National Geospatial-Intelligence Agency (NGA), in partnership with the Civil Applications Committee (CAC), of which NASA is a member, provides access to its immense archive of unclassified commercial high-resolution satellite data to non-DOD government agencies under terms of its NextView contract.

The NextView contract stipulates that the data can be used by all branches, departments, and offices of the U.S. Government. With appropriate approval and acknowledgement, the data can also be shared with NGO's, state/local governments, Intergovernmental agencies, as well as universities, and foreign governments if the use is in support of U.S. government interests. Goddard Space Flight Center has been tasked with providing an interface to these data for NASA investigations and their teams. The NGA imagery licensed under NextView are made available via their Web-based Access and Retrieval Portal (WARP).

Land-cover/land-use change in southern Vietnam through the lenses of conflict, religion, and politics, 1980s to present

Land Cover/Use

Decades of conflict, colonialism, growing population, and global agriculture commercialization have resulted in land-cover/land-use change (LCLUC) on multiples spatial scales throughout Southeast Asia. These changes have had a profound impact on the ethnic minorities, particularly in southern Vietnam. Vietnam has experienced significant political, economic, and environmental change since the 1950s and the end of colonialism and French Indochina. All LCLUC in Vietnam must consider the impacts of the Vietnam War and subsequent regional and internal conflicts, the establishment of the one party government under the Communist Party of Vietnam, recent market and trade liberalization, and a complex religious and sociocultural tapestry. This project focuses on the Đồng Tháp and An Giang Provinces of the Mekong Delta region, which are home to some of Vietnam's largest ethnic minorities, including Khmers and Cham people. They are also home to a uniquely Vietnamese form of Buddhism, Hòa Hảo, which figures importantly in the modern history and landscape of the region. Hòa Hảo emphasizes the connection of an individual to the land in a relationship that is intimately ethical, spiritual, and national. The advent of the satellite era enables studies of the physical changes on the environment but to fully understand the trajectory of landscape change it is necessary to incorporate the social and religious factors endemic to the region.

We propose to map the changes and model the future trajectory of LCLUC by incorporating a sociocultural framework in a spatial modeling environment for the Mekong Delta region of southern Vietnam, with humanistic and sociological studies combined with very high resolution LCLUC in the Đồng Tháp and An Giang provinces. This project will map all agricultural, forest, and urban LCLUC around the Tràm Chim National Park in Đồng Tháp Province, agricultural areas in both provinces, and the two cities of Cao Lãnh, Đồng Tháp Province and Long Xuyên, An Giang Province using Landsat and very high resolution (VHR) Digital Globe data from NASA Commercial Archive Data for years 1985 to 2018. In year 3, the VHR LCLUC mapping will extend to the entire Mekong Delta region. This project will utilize decision tree algorithms within a data science approach to mine the Landsat archive and WorldView-1, -2, and -3 data on the large computing capacity of the Advanced Data Analytics Platform (ADAPT) at NASA GSFC’s NCCS (http://www.nccs.nasa.gov/services/adapt). Using a mixed methods approach of historical documentation, in-country interviews, qualitative method of cultural vignettes, and quantitative methods of socioeconomic development pathways, we will extract historical, current, and future land cover/land use change trajectories and theories of change. Historical documents will be retrieved in-country and from the U.S. Library of Congress. In addition to socioeconomic data, this project will focus on the impact of conflict, religion, and political changes on LCLUC. These sociocultural and socioeconomic variables are important given the complex religious, ethnic, and economic tapestry of the region. These theories of change will be used to create scenarios of future LCLUC mapped to Boolean grids (Swetnam et al., 2010) and tested against a Markov chain approach. GIS models of future LCLUC will combine the remote sensing-derived products with the spatially explicit theories of change and other suitability variables within geostatistical weighted models and tested against the large region. Working with our in-country collaborator, open source web visualization and Atlas.ti KML linked file will be created to display historical, current, and predicted spatially-explicit 10-30 m resolution LCLUC of these two provinces. All data products created in the project will be shared in collaboration with the NASA SERVIR-Mekong project led by Collaborator Potapov.

Prototyping MuSLI canopy chlorophyll content for assessment of vegetation function and productivity

Land Cover/Use

Land cover use practices and the ongoing human activities and climate changes have significantly affected agricultural and forest productivity by imposing severe and novel combinations of multiple stresses on the natural ecosystems. There is a strong need to develop an approach for quantifying the spateo-temporal changes in vegetation condition and photosynthetic function at moderate ground resolution (20-30 m) across large regional/continental/global scales. In the spring of 2015 ESA’s Sentinel-2 (S-2) satellite joined NASA’s Landsat-8 (L-8) in providing moderate-resolution, multispectral measurements with global extent, therefore increasing the temporal resolution of such data. We propose to use the recently developed homogenized L-8 and S-2 (HLS) high-frequency time series to develop a new canopy chlorophyll content product, and to evaluate the seasonal changes in land cover chlorophyll content and associated productivity for key agricultural crops, grasslands and forested ecosystems.

The key objectives of the proposed effort are to: 1) using in a seamless fashion the dense time series of HLS, L-8 and S-2 images to develop algorithms for estimating canopy chlorophyll (Chl) content; and 2) generate robust workflows and produce high density time series of land cover Chl products for major vegetation cover types (crops, grasslands and forests).

Earlier studies using individual Landsat scenes (e.g. Landsats TM and ETM+) were not able to detect the early stages in vegetation damage. Those studies only partially accounted for variations in atmospheric conditions, terrain elevation and illumination. Using the improved spectral resolution of the HLS L-8 and S-2 data (narrower red edge bands and additional blue, near-infrared, short-infrared and thermal bands), complemented with very high resolution (2 m) World View images/triplets to characterize canopy variations and structural effects, we will produce dense time series of vegetation indices and Chl products sensitive to the fine changes in chlorophyll content for improved monitoring of agricultural crops and forest biomass production.

To generate robust workflows, the algorithms will be tested, refined and validated at established research areas and instrumented sites representing major agricultural crops and forested ecosystems. We will use the L-8 thermal bands (TIRS1) to quantify the effects from changes in Chl content and vegetation damage on agricultural and forest productivity, comparing the phenology trends in canopy Chl, TIRS1 and ecosystem primary production, as measured at the sites. By analyzing these coupled dense time series, we will provide
essential information about the major biophysical drivers of vegetation health and function.
This effort leverages the ongoing international collaborations between USA and European Union researchers, which will provide expertise and satellite imagery available at their organizations as well as field spectral data obtained from established sites with ongoing field data collections. The proposed work directly supports the goals of NASA’s LCLUC program, to further "develop the capability for periodic satellite-based inventories of land cover, and monitoring and characterizing land-cover and land-use change." This effort provides a significant step forward towards developing an approach and tools for evaluation of vegetation health and photosynthetic function at 30 m resolution, that will enhance our ability to identify the drivers and quantify the rates of land cover change for critical vegetation types across the globe. This enhanced capability will greatly improve the information available for timely management decisions that have potential to reduce the associated agricultural, economic and climate impacts of environmental and anthropogenic factors.

Physical Research Scientist

NASA - NASA Goddard Space Flight Center, Biospheric Sciences Laboratory

October 2010 - Present

 • Serves as a specialist that analyzes, interprets, and quantifies terrestrial disturbance impacts from natural events of fire, insect outbreak, and anthropogenic land-cover land-use change characterizing dynamics to identify and understand the distribution of sources and sinks of carbon in Northern Hemisphere ecosystems;
• Conducts physical science studies with remote sensing datasets obtained from space and complimented by airborne and ground-based measurements suitable for assimilation into numerical biogeochemistry simulations to develop empirical, theoretical, or experimental means to understand terrestrial disturbance processes relevant to ecosystem carbon cycling;
• Reviews, advises on, and conducts complex analysis of vegetation mapping integration into terrestrial carbon cycling simulations to quantify ecosystem responses to climate change, estimating trace gas fluxes from natural and human managed systems that are critical biophysical processes associated to vegetation productivity and ecosystem carbon storage;
• Documents and reports study results to the public and serves as authoritative source of consultation.

NASA Post-doctoral Fellow

Oak Ridge Associated Universities - NASA-GSFC

May 2010 - October 2010

 • Served as a specialist that analyzed, interpreted, and quantified terrestrial disturbance impacts from natural events of fire, insect outbreak, and anthropogenic land-cover land-use change characterizing dynamics to identify and understand the distribution of sources and sinks of carbon in Alaskan ecosystems;
• Documented and reported study results to the public.

Senior Research Scientist

Science Systems Applications Inc. - Greenbelt, Maryland

July 2001 - May 2010

 • Analyzed, interpreted, space based empirical measurements of human and natural disturbances to vegetation productivity and structure using multi-temporal multi-spectral satellite data to derive biophysical information to investigate trace gas emissions related to North America terrestrial carbon cycling;
• Planed, obtained, coordinated and implemented multi-temporal multi-spectral ground and satellite information for numerical model-data fusion of biogeochemistry simulations to quantify, improve, and discover the nature, magnitude, and distribution of terrestrial carbon sinks and sources that impact ecosystem carbon balance;
• Produced proposals and scientific manuscripts suitable for peer review and publication.

Applications Scientist

MDA Federal (Formerly Earth Satellite Corporation) - Gaithersburg, Maryland

April 1996 - July 2001

 • Conducted physical science studies through analysis of multi-temporal multi-spectral mapping of land-cover land-use change of vegetation and forest cover dynamics with orthorectified Landsat Thematic Mapper (TM) and Landsat Multispectral Scanner (MSS) remote sensing data;
• Coordinated and integrated development, oversight and scheduling of selection, acquisition, processing, and error estimation of global Landsat Multispectral Scanner (MSS) data for the GeoCover Ortho NASA project from U.S. and foreign ground stations.

American Association for the Advancement of Science, (AAAS)

2008 - Present

American Geophysical Union, (AGU)

2004 - Present

American Society for Photogrammetry and Remote Sensing, (ASPRS)

2004 - Present

Association for American Geographers, (AAG) Mid Atlantic Division

Treasurer 2006 - 2008

2004 - 2010

Advanced information systems to fill STV gaps: next-generation stereo+lidar fusion and sensor technology - NASA ESTO DECADAL SURVEY INCUBATION PROGRAM: SCIENCE AND TECHNOLOGY (NNH21ZDA001N-DSI) - Awarded: 2022-08-01

Dates: 2022-08-01  - 2025-07-31

Coverage: 0.1

The Impact of Investment on Irrigated Rice, Dryland Agriculture and Afforestation in Senegal using SAR and Optical Time-Series Imagery in Data Fusion Approaches

The Impact of Investment on Irrigated Rice, Dryland Agriculture and Afforestation in Senegal using SAR and Optical Time-Series Imagery in Data Fusion Approaches - NASA LCLUC - Awarded: 2021-01-01

Dates: 2021-01-01  - 2023-12-31

Coverage: 0.2

Towards a Warmer, Less Frozen Future Arctic: Ecosystem Impacts, Drivers, and Societal Responses Along Bioclimatic Gradients in Western Alaska - NASA Terrestrial Ecology ABoVE (NNH21ZDA001N-TE) - Awarded: 2022-05-01

Dates: 2022-10-01  - 2025-09-30

Coverage: 0.1

Mapping boreal forest biomass recovery rates across gradients of vegetation structure and environmental change

Mapping boreal forest biomass recovery rates across gradients of vegetation structure and environmental change - NASA Terrestrial Ecology ABoVE (NNH21ZDA001N-TE) - Awarded: 2022-05-01

Dates: 2022-10-01  - 2025-09-30

Coverage: 0.05