LCLUC 2019 Webinar Series

Presenter:  Ruth DeFries

Professor of ecology and sustainable development at Columbia University in New York. 

Tropical Deciduous Forests of South Asia: Monitoring Degradation and Assessing Impacts of Urbanization.
This project proposes to 1) develop remote-sensing methods to monitor forest degradation and regeneration in tropical deciduous forests of South Asia and 2) assess how off-farm employment associated with urbanization alters uses of forest resources that lead to degradation. While remote sensing to map deforestation are now welldeveloped, methods for mapping degradation (i.e. reduction in biomass or canopy cover) and regeneration are more complex and less mature. Moreover, considerable research has examined urban demands for agricultural commodities on deforestation. But the impact on forest resources of increasing urban migration and off-farm employment in South Asia, where rural populations rely heavily on forests for energy, fodder and incomegeneration from non-timber forest products, is not well understood. We will focus on a study region in central India, a globally important landscape for tiger conservation and home to millions of people whose livelihoods depend on forest resources. The region covers 250,000 sq. km. (approximately 8 percent of Indias land area). Remaining patches of tropical deciduous forests, the native vegetation, provide crucial connectivity for large mammals to travel between protected areas. Similar to other places in South Asia, rapid urbanization is generating off-farm employment that is changing resource use in rural households. Consequences for forest degradation and regeneration are not currently known. For example, increased income from off-farm employment could exacerbate grazing pressures through increased cattle ownership or alleviate pressures through reduced reliance on fuelwood. The project builds on our experience in the study region, prior work on remote sensing of forests, and household surveys on resource use in India. Initial efforts to map changes in biomass in the study region using L-band ALOS PALSAR and Landsat data show moderate success. Recently available radar (e.g., ALOS PALSAR 2, Sentinel 1 C-band) and optical (Sentinel 2, Quickbird and other very high resolution sensors) open opportunities to further develop methods for mapping forest degradation and regeneration. We will collect approximately 100 biomass measurements across the landscape to test multiple algorithms to predict biomass. We will interpret optical data to identify visible signs of degradation such as footpaths. The goal of the remote sensing analysis is to identify the data sources and algorithms to monitor forest degradation and regeneration efficiently, accurately, and with as minimal data processing as possible. The ultimate objective is to establish an operational monitoring system, maintained by Indian scientists, which will enable targeting efforts to reduce degradation, such as alternative livelihood projects, evaluate effectiveness of such efforts, and monitor progress towards reforestation goals. The social science aspect of the project will identify how rapid urbanization is changing demand for forest products and hence altering patterns of forest degradation and regeneration. We will develop a sampling strategy for 5,000 household surveys in 500 villages to cover the range of conditions across the study region. The surveys will quantify cattle ownership, dependence on fuelwood for energy, and collection of nontimber forest products in relation to off-farm employment and migration of family members to urban areas. We will test whether resource use is statistically related to employment of household members in urban areas (controlling for confounding factors). Results will indicate whether future urbanization is likely to alleviate pressures and lead to forest regeneration or to exacerbate pressures and lead to further degradation.

Presenter:  Karen Seto

Professor of Geography and Urbanization Science, Yale School of Forestry and Enviromental Research 

Urban Growth, Land-Use Change, and Growing Vulnerability in the Greater Himalaya Mountain Range Across India, Nepal, and Bhutan
Home to about 210 million people and extending over eight countries, the Hindu Kush Himalayan (HKH) region is at the confluence of two major trends that together are transforming one of the most dynamic mountain systems in the world. First, the region is a hotspot for four natural hazards: earthquakes, fires, floods, and landslides. Over the past few years, the HKH region has experienced a number of devastating natural disasters, including a 7.5 magnitude Pakistan-Afghanistan earthquake in 2015, a glacial lake outburst flood in northern Bhutan in 2015, floods in Uttarakhand in 2013 that left nearly 6,000 dead and more than 100,000 people trapped, and the 7.8 magnitude earthquake in Nepal in 2015, that killed more than 9,000 people and injured more than 23,000.
Second, the HKH region is rapidly urbanizing. Fueled by migration from rural areas, valleys and plains, the growth of religious, ecological and adventure tourism, and recent social unrest, towns and urban centers are expanding. Although the region is still predominantly agrarian, migration to urban centers is increasingly an important livelihood strategy for rural households, and non-farm income is an increasing component of household incomes. The growing urban population, an urbanizing economy, and associated land use and land cover changes are transforming the Himalayas. Yet despite the vulnerability of the region and its people, the 2015 Nepal earthquake highlighted the lack of accurate and up-to-date information about urban settlements in the region and those most at risk in this coupled social-environmental system.
The proposed research aims to fill these knowledge gaps by using multi-scale and multisource satellite data applied to novel and holistic vulnerability frameworks to answer five research questions about the HKH region:
1. How and where are urban settlements changing, and what are the associated land use and land cover changes with these changes?
2. What are the frequency, magnitude, and duration of the four dominant natural hazards—earthquake, fire, flood, and landslides, and how do they vary over time and space?
3. What is the sensitivity of the socio-economic system to different stressors?
4. Where are urban settlements most vulnerable and to what stressors are they most vulnerable?
5. What explains differences in the vulnerability of urban settlements across the HKH region?
The two primary goals of the proposed research are to 1) characterize and quantify LCLUC associated with urban settlement change, and 2) assess the vulnerability of these urban settlements to hazards. The project will be undertaken at two spatial extents. A 30m resolution analysis will be conducted for 41 contiguous scenes for the entire Landsat TM archive (1984-present) covering an area of approximately 1.289 million km. This wall-to-wall approach using the entire Landsat archive is a marked departure from most other urban studies in the region that focus solely on a few capital or large cities and their immediate surroundings. To examine the accuracy of the Landsat analysis, we will use 2.5m resolution imagery to analyze LCLUC in three sites. Expected project results and benefits include: New remote sensing algorithms to characterize changes in urban settlements and associated LCLUC in high mountain regions of South Asia. Empirical estimates of urban LCLUC, including transport infrastructure and the built environment in the HKH region of India, Bhutan, and Nepal. Improved scientific understanding of the spatio-temporal patterns of land use change in the HKH region.

Presenter:  Liping Di

Professor , Geaorge Mason University 

Understanding Changes in Agricultural Land Use and Land Cover in the Breadbasket Area of the Ganges Basin 2000-2015: A Socioeconomic-Ecological Analysis
In South Asia, agriculture faces the remarkable challenge of feeding a growing population, projected to increase 43.8% by 2050, with extremely limited per capita arable land. Adding to the complexity of the challenge is the rapid economic development and urbanization over recent years, which competes for land with the agricultural sector and threatens the remaining forest cover. The Ganges Basin is the breadbasket of South Asia. Its agricultural outputs feed almost one-tenth of the world population. The densely populated region is home to numerous cities that have undergone rapid expansion in recent decades. Forest and wetland ecosystems are under increasing threats from intensified agricultural land use and pressure from urban spatial expansion. South Asia plays a central role in the global effort (such as the United Nations Sustainable Development Goals) to achieve sustainable development and food security. A deeper understanding of the drivers and socioeconomic and ecological impacts of land cover and land use change (LCLUC) in the basin is essential to better inform land use policies. This proposed project will combine remote sensing and geographic information system (GIS) techniques with an integrated modeling approach to 1) identify and assess LCLUC that occurred in the major agricultural area of the Ganges Basin (including India and Bangladesh) during 2000; 2015; 2) identify and quantify the primary socioeconomic drivers of LCLUC; 3) develop scenarios of future LCLUC (up to 2030) based on projected changes of key drivers; 4) evaluate the main impacts of LCLUC on key indicators of food and nutrition security, income, ecosystem services, land degradation, and resilience; and 5) disseminate research findings and reach out to stakeholders to ensure that the research helps to inform sustainable development strategies and responds to the needs of policy makers. In this project, LCLUC includes changes in agricultural land cover/use at three levels: 1) land use change from agricultural to non-agricultural use, or vice versa; 2) change in cropping systems, such as from rice to maize; and 3) change in the intensity of agricultural land use. The integrated modeling framework includes three modeling components: 1) a global agricultural partial equilibrium model, 2) a spatially explicit econometric land use model, and 3) an ecosystem services evaluation model. This research will enable us to answer some of the key scientific questions related to LCLUC in the basin: What are the dominant LCLUCs, including agricultural intensity and cropping system changes, in the breadbasket region of the Ganges Basin during 2000’2015? What are the major socioeconomic drivers of LCLUCs during 2000’2015 and how did economic development, population growth, and the accompanying structural transformation such as urbanization affect the changes? What are the impacts of LCLUCs on key indicators of food and nutrition security, income, ecosystem services, land degradation, and resilience? This proposed project will make two novel contributions to NASAs LCLUC program. First, it will develop state-of-art remote-sensing techniques to identify level-2 and level-3 LCLUC in developing country context where farm and field sizes tend to be small. Second, this project will employ economic theories and methods, particularly an integrated modeling framework, to identify and quantify major drivers and impacts of LCLUC in the study area. This proposed project will contribute to the scientific knowledge base regarding integrated dynamics of LCLUC and socioeconomic systems at regional, national, and global scales, enhance evidence-based and science-guided decision-making concerning food security and sustainable agricultural development, and support the United Nations 2030 Sustainable Development Goals.

Presenter: Randolph Wynne

Professor, Virginia Polytechnic Institute and State University

Spatiotemporal Drivers of Fine-Scale Forest Plantation Establishment in VillageBased Economies of Andhra Pradesh
The Indian state of Andhra Pradesh, our study area, has decreased in overall forest cover in recent years, with concomitant (though not commensurate) increases in forest plantation area, largely through conversion of degraded and existing agricultural land. Unfortunately, however, accurately mapping forest plantations in India using remotely- sensed data has been difficult because: (1) many plantations are small relative to moderate resolution earth resource satellite data, (2) newly established plantations are very difficult to identify, and (3) the surrounding cropland area is very variegated in both time and space. A concomitant socio-economic issue is the relatively unknown incentives and risks associated with establishing plantations within the broader land use context at the decision maker level in this region. Leveraging forest industry partnerships and strong extant approaches pioneered by our team, our overall goals are to (1) improve the accuracy and precision by which forest degradation and plantation establishment can be remotely-sensed using data from ResourceSat-1, Landsat, SPOT, and/or RapidEye; and (2) determine the most predictive local, regional, village, and household based drivers of plantation forest establishment (the social science aspect of the proposed study). Our proposed research is particularly responsive to the solicitation, as it (1) couples remote sensing observations (using both NASA and synergistic non-NASA assets) from which land-cover can be derived with research on the human dimensions of land-use change, (2) is explicitly focused on the South Asia geographic region of interest, (3) improves the detection, monitoring, and predictions of land cover and land use change in the region, (4) attributes land cover and land use changes to their primary causes, and (5) provides a formal means and rigorous method for evaluating the socially best land use mix over time as the region develops. The primary expected outcomes are as follows: (1) improvements to algorithms from which both discrete and continuous land use and land cover change variables can be remotely-estimated in this tropical, fine-scale, temporally dynamic, and spectrally variegated landscape, and (2) an empirical realization of forest plantation establishment in village-based economies where smallholders establish forest plantations on previously degraded lands for both timber and non-timber use, using a unique data set developed over time and space through the period covered by the proposal. Deliverables will include, but not be limited to, the following: (1) a map of plantations, natural forests, degraded forests, and primary non-forest land uses in East and West Godavari, (2) a system of equations resulting from the econometric analysis, with a rent function describing the net returns from each land use option for a household, and a response model of the decision to establish forest plantations on existing and new land as a competing land use with other uses such as agriculture and grazing. Our tentative schedule is as follows: Year 1: (1) develop econometrics survey instrument, (2) use high-resolution remotely-sensed data in concert with in situ reference data to (a) develop tree canopy cover dependent variable distributions (b) map plantations and other key land uses in East and West Godavari circa 2016, (3) use resulting map and ancillary data to develop sampling strata for survey, (4) distribute and collect survey instruments. Year 2: (1) refine tree canopy cover (both static and dynamic) estimation approaches using Landsat data, (2) analyze survey data; develop econometric models. Year 3: (1) map canopy cover and change in the region, (2) assess how property rights risks and future market opportunities for competing uses affect forest plantation establishment, (3) quantify the errors associated with non-integrated land use change modeling. Overall, our study has strong potential to help enrich LCLUC science in South Asia.

Presenter: Aditya Singh

Assistant Professor, University of Florida

Landscapes In Flux: The Influence of Demographic Change and Institutional Mechanisms on Land Cover Change, Climate Adaptability and Food Security in Rural India
While there has been a considerable reduction in the number of undernourished people in the past two decades, India still has among the largest malnutrition rates in the world. Increasing pressures from population growth and urbanization have subsequently affected land use patterns in India, with as much as 36.6% of all land in India degraded. Climate change will likely exacerbate these issues through additional stresses on food production. Regionally, a range of socioeconomic factors also play a role, such as: lack of availability of and access to resources, land degradation, food insecurity and landlessness. Geographically differentiated strategies are needed to address these issues, but current strategies are severely hampered by the paucity of spatially-explicit information on food security and agriculture. This information is crucial for early detection of trends and to disentangle the complex relationships between food security and land use. Of critical need are spatially-explicit data on the factors that define dimensions of food security, and methods that allow the combination of these data into holistic synthetic indicators that explain causal factors in an integrated manner. Identifying, estimating and mapping spatial variations in the proportional strengths of these interrelationships will help identify the factors influencing food security and eventually land-use and land-cover change in rural as well as peri-urban areas. The specific objectives of this proposal are to: 1) generate spatially downscaled data of key demographic and socio-economic indicators that putatively define dimensions of food security in India, 2) use a hypothesis-driven approach to integrate economic, social, policy, infrastructural, and behavioral facets of food security into a holistic modeling framework, and, finally to 3) assess land cover change as an emergent outcome of patterns of socio-economic, demographic and policy instruments at local to regional scales. To do this, we will use small area estimation techniques to spatially disaggregate household-scale data on critical demographic, socioeconomic and food security indicators to the village scale. Subsequently, using a structural equation modeling framework, we will integrate indicators of food security with extant socioeconomic and demographic patterns, indicators of climate adaptability and metrics of infrastructural and policy instruments. Maps of latent vectors of the SEM will allow the first-ever spatialized representation of the combined effects of institutional support, accessibility to markets and extension services on regional indicators of poverty and malnutrition. Further, we will develop a generalized methodology for mapping land cover and producing probabilistic pixel-wise maps of classification uncertainties. We will eventually combine land cover change probabilities with indicators of food security derived from structural equation models to assess the influence of food security indicators on patterns of land cover change. These analyses will provide the first-ever assessments of the relative strengths of drivers of land cover change in the study regions at local to regional scales. The proposed research directly addresses NASAs high-priority science goals with a central focus on Land Cover Land Use Change science within the larger Carbon Cycle and Ecosystems program. In addition, the proposal directly addresses the influence of socio-economic drivers on land cover change. By developing a strong socio-ecological context to all our study sites and analyses, we will ensure the interdisciplinary application of space-borne technologies to help address issues of high societal relevance. The proposed research is therefore directly responsive to the NASA LCLUC program themes: detection and monitoring of change, predictive land use modeling, climate variability and change, and drivers of change and food security.
 

Presenter: Jefferson Fox, Kaspar Hurni

Center For Cultural And Technical Interchange Between East and West

The Agrarian Transition in Mainland Southeast Asia: Changes in Rice Farming - 1995 to 2018
This project responds directly to the solicitation for LCLUC studies in Southeast Asia by examining how the region is responding to simultaneous loss of agricultural labor and intensification of rice production. Major project objectives include: 1) Build a comprehensive multi-resolution, satellite image-derived database to characterize variability across and long-term changes within regional rice production systems; 2) Model current and past rice production under changing socio-economic and environmental conditions; 3) Use national population and agricultural censuses and other spatially-explicit secondary datasets compiled for sub- district units to quantify how changing conditions are correlated with changes in rice production systems through time; and 4) Conduct field interviews at selected sites to develop a place-based understanding of how rice farming is being revolutionized by changing demographics, economic opportunities, and technological innovations. We will explore these objectives for the major rice producing areas of four Mainland Southeast Asia (MSEA) countries (a total of six rice producing regions) between 1995 and 2018. The four countries and six regions include: 1) Vietnam (Red River and Mekong River Deltas), 2) Thailand (Northeast and Central Regions), 3) Laos (Savannakhet Province), and 4) Cambodia (Battambang Province). We will quantify changes in rice production systems between 1995 and 2018. As a means of quantifying long-term landscape dynamics in the persistently clouded study area, we will use an assemblage of complementary, cloud-resilient remote sensing analytical methods. First, we will classify Sentinel 2 SAR time series data (2014-2018) through an unsupervised rule-based clustering algorithm to differentiate stable standing water from flooded rice paddies to map locations and timing of rice production. Second, we will apply the Noise Insensitive Trajectory Algorithm (NITA) on Landsat (1995- 2018) and Sentinel 2 (2015-2018) time series data to quantify and map sub-annual changes in timing and pattern of rice production across our six study regions. NITA models land cover dynamics for every satellite image pixel across all available image dates and is the first all-available-images time series algorithm specifically designed to process data suffering from signal degeneration due to atmospheric effects. We will then input satellite-derived measures of area under rice production to the CSM-CERES-Rice model to estimate plot-level as well as regional annual rice yields. We will examine quantitative relationships between rice production, physiographic variables, and socioeconomic data gathered from agricultural and national censuses. A regression forest relating socioeconomic and physiographic variables to change in rice production systems will be used to identify the most significant predictors of change in each study region and across the study area. Finally, to understand how changes in labor dynamics and increasing demands for off-farm employment alter processes associated with rice production in land preparation, planting, weeding, harvesting, and the number of crops grown per year, we will conduct semi-informal interviews with key informants and survey 100 households in each rice growing region (total of 600 households). The project’s significance to NASA lies in its improved, multi-sensor approach for mapping changes in rice production systems-a change in land use rather than land cover, its use of novel cloud-resilient LCLUC monitoring approaches, and its integration of regional and local-scale perspectives of conditions that underlie observed changes to rice production systems (e.g., urbanization or industrialization). The knowledge generated by the proposed research will improve understanding of the social and ecological transformations affecting MSEA rice production and broadly advance globally-relevant theory on agriculture adaptation and change. 

Presenter: Laixiang Sun

University of Maryland, College Park

Forest Change and Oil Palm Expansion in Southeast Asia: Historical Patterns, Socioeconomic Drivers, and Future Projection

Palm oil is currently the most consumed edible oil in the world. According to USDA, the worldwide production of palm oil has increased from 15 million tons in 1995 to 63 million tons in 2016. Indonesia and Malaysia have been the biggest suppliers of palm oil since 1966, with the dominant share of 85% in 2016. The plantation of oil palm (monoculture or mixed) amounts to 14 million ha in Indonesia and 7 million ha in Malaysia, which account for 62% and 84% of the total plantation areas in each of these two countries, respectively. The FAO land-cover data show that more than 55% of oil palm expansion during 1990-2005 in these two countries occurred at the expense of natural forests, and the remaining occurred mainly at the expense of existing agricultural land. In an increasingly health-conscious world, global demand for palm oil is bound to increase in the near future as consumers shift towards consumption of vegetable oils containing low trans-fat. An additional driving force for rising demand on global palm oil market is the growing bio-fuel blending demand posed by climate change concerns. A 2015 study of Grand View Research Inc. indicates that the global palm oil market demand is likely to increase to 128 million tons in 2022, with an annual growth rate of 7.5%. The above discussion indicates that there are tough challenges for policymakers and other stakeholders to b