Space Enabled Presents at the 2023 American Geophysical Union Conference


AGU 2023

AGU 2023

AGU is an international, nonprofit scientific association whose mission is to promote discovery in Earth and space science for the benefit of humanity. Every year, AGU Fall Meeting unites over 25,000 attendees from 100+ countries in the Earth and space sciences community to discuss findings, connect scientists from around the world, advance their profession and connect over the passion for the impact of science. This year, members of Space Enabled will be presenting several papers and participating in panels and talks. Danielle Wood, Katlyn Turner, Sharif Islam, Ufuoma Ovienmhada, and Priscilla Baltezar will all give talks at AGU 2023. Danielle and Ufuoma are also co-hosting sessions.

Monday, December 11:

Sharif Islam

TitleShifting Coastlines, Human Displacement, and Erosion Risk Vulnerabilities in Bangladesh

Time: 8:30am-12:50pm PST

Place: Poster Hall A-C - South (Exhibition Level, South, MC)


Coastal Bangladesh has one of the highest rates of coastal shoreline erosion in the world. Erosion causes numerous problems including population displacement and loss of houses and household assets, productive lands, infrastructure, and livelihoods. Due to climate change and associated sea level rise, it is expected that erosion intensity and magnitude will increase in the future. This study aims to advance the understanding of the drivers of resident’s erosion risk predictions using spatially explicit measures of coastal erosion risk derived from Landsat satellite imagery and a random survey of residents living along the eastern bank of the Meghna estuary located in the coastal Bangladesh. Prior empirical work has documented that this region has experienced extreme rates of erosion since 2000. With an aim to assess the drivers of accurate risk predictions, this study uses Logistic Regression (LR) modeling to examine the roles of demographic, economic and location variables as predictors of accurate or inaccurate perceptions of subsequently measured erosion occurrence. Results suggest that the accurate prediction of erosion risk by surveyed households is significantly influenced by the physical location of the house. Further, predictors such as proximity to the coast and whether respondents’ house locations were unprotected by a recently constructed revetment are strongly associated with the accurate prediction of coastal erosion risk. This study highlights the critical role socio-economic and locational attributes play in risk perception at the household scale. Findings of this study inform the need to raise awareness for better planning and management of coastal areas and thus resilience of coastal populations to a changing climate. Research findings can also assist in the development of associated mitigation and adaptation measures that better incorporate community perceptions of risk.

Ufuoma Ovienmhada

Title: Spatial Pattern of Land Surface Temperatures and PM 2.5 in U.S. Prison landscapes using satellite-derived data

Time: 10:35-10:45pm PST

Place: 2016 - West (Level 2, West, MC)


In recent years, journalists and researchers have elevated a pattern of prison landscapes being exposed to environmental hazards such as air pollution, poor water quality, proximity to hazardous waste facilities, and inadequate mitigation in extreme weather conditions - a pattern of environmental injustice frequently referred to as “prison ecology”. This intersection of mass incarceration and environmental harm has been largely understudied in ecological terms or at scale.

We partner with the Campaign to Fight Toxic Prisons (CFTP), a national collective — composed of environmental justice organizers, formerly and currently incarcerated folks and their loved ones — organizing resistance at the intersection of incarceration and the environment. In this presentation, we highlight findings from in-depth interviews with CFTP and other organizers on opportunities to leverage satellite remote sensing data for advocacy and environmental justice. These findings are used to inform the development of two geospatial datasets evaluating Land Surface Temperature measurements and PM 2.5 concentrations in carceral landscapes across the U.S. We also present progress on a co-designed geographic information system that integrates satellite and socioeconomic data to respond to data needs in decision-making workflows for prison ecology activist organizations. These workflows include efforts to improve material conditions of incarcerated people exposed to environmental hazards and longer-term efforts to affect policy and governance of controversial prison projects in toxic landscapes. The findings can be used by community organizers, policy makers, and anyone seeking to advocate for environmental justice for a population of people that tend to sit at the margins of fights for human rights.

Danielle Wood

TitleUsing Environment-Vulnerability-Decision-Technology (EVDT) to Transfer Science to Action for Biodiversity Management in West Africa

Time: 11:30am-11:40am PST

Place: 3009 - West (Level 3, West, MC)


Governments at local and national levels are working to simultaneously manage the concerns of income inequality, climate change and biodiversity loss. These government teams need evidence-based tools to inform their policies and practices. Satellite based data, along with other sources of earth observation such as models, in-situ sensors, airborne sensors and field work, can be vital sources of information to support policies. Many of these policies are guided by efforts to improve performance based on the Sustainable Development Goals, especially Goals number 15 (Biodiversity on Land) and 13 (Climate). One challenge facing government teams in agencies focused on census data collection, forestry, agriculture, coastal management and water resources is that satellite data requires specialized expertise to apply to policy. Further, environmental data is more effective to inform policy when it is combined with socioeconomic and policy data. Addressing the SDGs requires multidisciplinary information combined in a manageable format. To enable such an approach, the Space Enabled Research Group at MIT developed the Environment-Vulnerability-Decision-Technology (EVDT) Integrated Analysis Framework. This tool provides a series of steps to iteratively design tools to package environmental, socioeconomic, policy and engineering information into a centralized visualization setting (usually a web-based map). Once an EVDT application is developed, multidisciplinary teams collaborate from a common base of information. This talk shows how such an approach provides knowledge derived from earth observations to support decision making in West Africa. One case study highlights how methods for land use classification apply machine learning to analyze medium resolution satellite data. The analysis achieves high accuracy and repeatable findings. The team next designed methods to facilitate adoption of the approach by agencies within the government of Ghana for use in national SDG#15 accounting. Similarly, the team is transferring methods to apply satellite data to track water hyacinth for use by local government science teams and an environmental company in Benin. In both cases, the team uses EVDT as a guide for understanding the needs of stakeholders and guiding the selection of data and visualization methods.

Ufuoma Ovienmhada

TitleSatellite Data for Environmental Justice: A Scoping Review

Time: 2:14-2:22pm PST

Place: 2014 - West (Level 2, West, MC)


In support of the environmental justice (EJ) movement, researchers, activists and policymakers often use environmental data to document evidence of the unequal distribution of environmental burdens and benefits across demographic and socioeconomic characteristics. Numerous limitations, such as spatial or temporal discontinuities, exist with commonly used data measurement techniques, which include ground monitoring and federal screening tools. Satellite data is well poised to address these gaps in EJ measurement and monitoring, however little is known about how satellite data has advanced findings in EJ or can help to promote EJ through interventions. This review 1) explores trends in study design, topics, geographic scope, and satellite datasets used to research EJ, 2) synthesizes findings from studies that use satellite data to characterize disparities across socio-demographic groups for various environmental categories, and 3) captures how satellite data are relevant to policy and real-world impact. We retrieved 81 articles that applied satellite data for EJ research in the United States from 2000 to 2022. The majority of the studies leveraged satellite data to identify socio-demographic disparities in exposure to environmental risk factors, such as air pollution, and access to environmental benefits, such as green space, at wider coverage and with greater precision than previously possible. These disparities in exposure and access are associated with health outcomes such as increased cardiovascular and respiratory diseases, mental illness, hospitalizations and mortality. Research using satellite data to illuminate EJ concerns can contribute to efforts to mitigate environmental inequalities and reduce health disparities. Satellite data for EJ research can therefore support targeted interventions or influence planning and policy changes, but the review reveals that significant work remains to facilitate the application of satellite data for policy and community impact.

Tuesday, December 12:

Priscilla Baltezar (co-author)

TitleInitial Co-Design of Earth and Climate Data Tools for Learners: Towards a Framework and Requirements

Time: 8:33-8:43 PST

Place: 203 - South (Level 2, South, MC)


The NASA Science Activation (SciAct) program connects NASA Science with diverse learners of all ages and seeks to increase learners’ active participation in the advancement of human knowledge. We will present the early findings from an MIT-led project that has partnered with a subset of funded teams from NASA’s SciAct program, all with a deep history of working with NASA Earth and Climate data for educational goals, to identify and characterize a concrete, discrete set of “last mile” obstacles that, if overcome, might provide the kind of data accessibility and use imagined in both NASA SciAct’s Objectives and NASA’s Equity Plan.

Using a participatory methodology known as co-design, this project centers the voices and needs of the audiences—in this case the educational needs of SciAct teams and learners—from the very beginning of the design process. The teams that we have partnered with for this project represent varied regional contexts and diverse learner populations across the nation. This diversity enables us to efficiently capture a broad range of educational use-cases that will result in more generally accessible and useful tools and design solutions. Collectively, the educational stories from these SciAct teams inform a developing framework that defines a set of general and abstract challenges for learning and data tool design that can be tackled by instructional designers working collaboratively with software designers and developers and NASA data producers and Subject Matter Experts (SMEs). In this presentation, we will share the themes and findings from our needs assessment interviews. We will also discuss the initial framework to represent identified needs and uses, informed by peer-reviewed literature, of effective application of Earth and climate data and tools in learning contexts.

This ongoing project will ultimately result in design specifications and recommendations for tools that improve the integration of Earth and Climate data into educational programming based on the known requirements of learners and educators.

Wednesday, December 13:

Priscilla Baltezar (eLightning Presentation)

TitleMulti-sensor Change Detection and Machine Learning for Invasive Water Hyacinth Identification in Lake Nokoué, Benin

Time: 8:30am-10:00am PST

Place: eLightning Theater I, Hall D - South (Exhibition Level, South, MC)


The water hyacinth is an invasive plant species that has spread throughout coastal West Africa, reducing biodiversity, clogging transportation networks and impacting fishing activities. Green Keeper Africa (GKA), an enterprise located in Benin, works to harvest around Lake Nokoué and repurpose the plant. The MIT Media Lab Space Enabled Research Group is partnering with GKA to design an online observatory that uses satellite data, drone imagery and ground based sensor data to map the location of the water hyacinth over time and publish this information for government, private and public users. The observatory also includes information about Acadja, a traditional fish farming practice that is related to water hyacinth dynamics and management initiated by Beninois fishing people. This project responds to Sustainable Development Goal 15.8 pertaining to the management of invasive species. This presentation will showcase the results of a random-forest machine learning model applied to classify water hyacinth and Acadja on Lake Nokoué for a period in 2022. We compare the accuracy of this model to a more traditional NDVI thresholding approach that leverages the spectral differences in water hyacinth compared to other vegetation. We also apply both methods to time-series data to evaluate spatial changes in water hyacinth extent over time. The water hyacinth and Acadja analyses have business and regulatory implications for Green Keeper Africa and management entities such as the Benin National Institute of Water, the Ministry of the Living Environment and Sustainable Development, the Ministry of Agriculture, Livestock and Fisheries, and the Ministry of Water and Mines. Green Keeper Africa can use the information presented about water hyacinth to improve their harvesting practices to stay in line with their priorities for profit and reduction of socioeconomic and ecological impacts. The Ministries can use this information to pursue new policies and practices that support SDG 15.8 and related SDG targets.

Danielle Wood

Title: Supporting Drought Management in Angola using Integrated Modeling with the Environment-Vulnerability-Decision Making-Technology (EVDT) Framework

Time: 10:45-10:55 PST

Place: 3020 - West (Level 3, West, MC)


An international team from Angola and the United States seeks to improve the use of satellite-based Earth Observation as an input to a Drought Decision Support System to inform the response to drought and floods in southern Angola. Specifically, the government of Angola needs to determine if interventions are delivered to the regions in which residents face high vulnerability. We define vulnerability based on the sensitivity, exposure and adaptive capacity of populations to drought. One long term policy intervention began in November 2019; the Angolan Ministry of Energy and Waters initiated the construction of a new system for catching, pumping and transporting water from the Cunene River as one of several measures to respond to the drought and flood cycle. Short term policy responses include using trucks to deliver emergency water and food. A Drought Decision Support System is proposed that can assist drought response. The project applies the foundational framework called “Environment-Vulnerability-Decision-Technology” (EVDT), which allows a structured process to create decision support systems for scenarios impacted by environmental state, human experience, policy decisions and the use of technology. The Environmental analysis applies NASA’s Soil Moisture Active-Passive data to estimate soil moisture across all of Angola. The Vulnerability analysis applies geospatial datasets that provide insight into the socioeconomic factors that influence the sensitivity and adaptive capacity of communities. The Decision analysis applies a Systems Architecture Framework to collect data from stakeholders in Angola regarding the factors that influence drought response policies. The Angola Drought Decision Support System contributes to a capability for the country’s space agency to provide space-based data to support government policy for water management. The presentation shows examples of combining these various perspectives using the EVDT framework.

Katlyn Turner

TitleSystems architecture as a tool for co-creative development of decision support systems: Angolan drought and climate resilience

Time: 2:10pm-6:30pm PST

Place: Poster Hall A-C - South (Exhibition Level, South, MC)


This project seeks to use a mixed methods engineering + social science approach to address environmental and climate resilience in southern Angola. In order to inform the response to drought and floods in southern Angola, a Drought Decision Support System is being developed. Southern Angola experiences periodic drought, particularly in the Cunene province. This drought can be severe and has a great impact on the country’s ability to provide services and infrastructure to its citizens, as well as on Angola’s economic impacts. The Drought Decision Support System seeks to improve the use of technical tools and infrastructure in Angola by combining several aspects: satellite-based Earth Observation data, the social vulnerability index, and systems architecture analysis in order to support the government of Angola’s decisions for drought intervention, planning, aid, and mitigation. Systems architecture is a framework that allows for a sociotechnical system to be analyzed descriptively and qualitatively. Systems architecture frameworks are used to support and understand both social and technical aspects contributing to organizations and technologies. Systems that work to try and create a decision support process–e.g. around natural disaster management, climate resilience, and ecosystem services– benefit from the process of using systems architecture, in particular when the work involves many stakeholders, power dynamics, communities, and organizations. Although earth science tools often have technical rigor, a rigorous approach to sociotechnical aspects of issues such as climate management is needed in order to create successful, inclusive, and durable solutions. One important aspect of systems architecture as a method is the process of co-creation, community science, and stakeholder analysis that helps earth scientists, designers, and policymakers understand a particular physical environment as well as the social environment surrounding a given environmental or infrastructural project. This work discusses this overall project to create a Decision Support System for management of drought in Angola, with a particular focus on the methodology of systems architecture analysis and stakeholder engagement as a tool to develop decision support systems for environmental and climate resilience.

Ufuoma Ovienmhada (Session Chair)

Title: Geospatial Data Applications for Environmental Justice II Oral

Time: 2:10-3:40 PST

Place: 2020 - West (Level 2, West, MC)


Environmental determinants of health are often unevenly distributed across populations, leading to disproportionate exposures and health risks among marginalized communities - a pattern commonly referred to as environmental injustice. Earth observing satellites, novel algorithms, low-cost monitors, mobile sensors, models, and citizen science monitoring are increasingly providing high spatial resolution and more complete geographic coverage of environmental conditions. This unprecedented data availability is opening new avenues for mapping and analysis to characterize disparities in the distributions of environmental stressors between population subgroups that have different ethnic, racial, socioeconomic, gender, immigration status, incarceration status, and other characteristics. This session will feature research that develops and applies these novel geospatial technologies and data for understanding and addressing environmental injustice for affected communities.

Thursday, December 14:

Ufuoma Ovienmhada (Poster Session Host)

TitleGeospatial Data Applications for Environmental Justice I Poster

Time: 8:30am-12:50pm PST

Place: Poster Hall A-C - South (Exhibition Level, South, MC)


Environmental determinants of health are often unevenly distributed across populations, leading to disproportionate exposures and health risks among marginalized communities - a pattern commonly referred to as environmental injustice. Earth observing satellites, novel algorithms, low-cost monitors, mobile sensors, models, and citizen science monitoring are increasingly providing high spatial resolution and more complete geographic coverage of environmental conditions. This unprecedented data availability is opening new avenues for mapping and analysis to characterize disparities in the distributions of environmental stressors between population subgroups that have different ethnic, racial, socioeconomic, gender, immigration status, incarceration status, and other characteristics. This session will feature research that develops and applies these novel geospatial technologies and data for understanding and addressing environmental injustice for affected communities.

Danielle Wood (Co-Author for Poster Presentation by Stewart Isaacs)

Title: Quantifying Natural and Anthropogenic Aerosol Impacts on Decentralized Solar in West Africa

Time: 2:10pm-6:30pm PST

Place: Poster Hall A-C - South (Exhibition Level, South, MC)


Decentralized solar is an emerging strategy for advancing modern energy access among rural populations in the West African region. However, ground-level solar irradiance can be reduced by local aerosol sources of both natural and anthropogenic origin. Airborne dust from the nearby Bodélé Depression and aerosols from industrial and agricultural emissions can negatively impact local solar energy systems, but the full extent of the effects of each is unclear.

In this work, we assess aerosol spatial and temporal distributions and variability from ground-based, reanalysis, and satellite-derived datasets. We produce a multi-year high-spatial resolution climatology of aerosol optical depth (AOD), evaluate the accuracy of AOD estimates within the datasets, and determine the contribution of dust and anthropogenic aerosols to the total regional aerosol load. We then estimate resulting irradiance reductions and their effects on solar energy applications.

We find that the largest aerosol loads are dust dominated during the annual Harmattan season. Aerosol loading can reduce global horizontal irradiance by up to 40%, which is equivalent in magnitude to the cloud-induced reductions observed during this season. We show AOD has the largest temporal variability in the inland locations of the region, which is driven by frequency of local dust events. Finally, we demonstrate decentralized solar systems are at risk of underperforming in reliability, depending on the accuracy of the aerosol dataset used during sizing.

Our results support the design of more reliable of solar systems in West Africa while providing insights to policy makers on the impact of reducing combustion emissions on regional solar power output.

Friday, December 15:

Ufuoma Ovienmhada (Town Hall Host)

Title: TH53J - The Role of the Geosciences in Environmental Justice: Perspectives from Community Leaders on Purposeful Science

Time: 1:00pm to 2:00pm PST

Place:  2016 - West (Level 2, West, MC)

Abstract:  GeoHealth combines earth, environmental, and health sciences across disciplinary boundaries to identify environmental and health injustices, examine societally relevant scientific questions, and collaborate in and with frontline communities for solution-oriented decision making at all scales. A core mission of the GeoHealth Section is to enhance AGU members’ engagement in community-engaged and community-led transdisciplinary research, which is key to conducting purposeful science. As part of this effort, we must look to community leaders to understand how they view the role of science in their advocacy and what makes for positive community engagement.

In this town hall, Environmental Justice community leaders will share their perspectives on the role of the geosciences in their advocacy and how scientists can best collaborate with them. Following short talks (8 mins), the panelists will engage in Q&A regarding their experiences in conducting or seeking geospatial data, citizen science, and more generally in working with underserved communities.

Danielle Wood (Presentation by Abigail Barenblitt as lead Co-Author)

Title: The large footprint of small-scale artisanal gold mining in Ghana (Invited)

Time: 4:50pm-5:00pm PST

Place: 2018 - West (Level 2, West, MC)


Gold mining has played a significant role in Ghana's economy for centuries. Regulation of this industry has variedover time and while industrial mining is prevalent in the country, the expansion of artisanal mining, or Galamseyhas escalated in recent years. Many of these artisanal mines are not only harmful to human health due to the useof Mercury (Hg) in the amalgamation process, but also leave a significant footprint on terrestrial ecosystems, degrading and destroying forested ecosystems in the region. In this study, the Landsat image archive availablethrough Google Earth Engine was used to quantify the total footprint of vegetation loss due to artisanal goldmines in Ghana from 2005 to 2019 and understand how conversion of forested regions to mining has changedover a decadal period from 2007 to 2017. A combination of machine learning and change detection algorithms were used to calculate different land cover conversions and the timing of conversion annually. Within thestudy area of southwestern Ghana, our results indicate that approximately 47,000 ha (⨦2218 ha) of vegetation were converted to mining at an average rate of ~2600 ha yr−1. The results indicate that a high percentage(~50%) of this mining occurred between 2014 and 2017. Around 700 ha of this mining occurred within protected areas as mapped by the World Database of Protected Areas. In addition to deforestation, increased artisanal min-ing activity in recent years has the potential to affect human health, access to drinking water resources and food security. This work expands upon limited research into the spatial footprint ofGalamseyin Ghana, complementsmapping efforts by local geographers, and will support efforts by the government of Ghana to monitor deforestation caused by artisanal mining.

Related Content