We look forward to welcoming you to the Google Booth in the Exhibit Hall, in Moscone North

Do stop by for some of the talks at our booth, listed below. Or come over and meet the Googlers who will be there to talk about our products and programs, discuss how you might use our tools to enhance and communicate your work, or to answer questions.

Google's booth will be open when the Exhibit Hall in Moscone North is open:

• Monday Dec 9, 6:00 PM - 8:00 PM - AGU Icebreaker Reception
• Tuesday Dec 10, 10:00 AM - 5:00 PM
• Wednesday Dec 11, 10:00 AM - 5:00 PM
• Thursday Dec 12, 10:00 AM - 5:00 PM
• Friday Dec 13, 9:30 AM - 1:30 PM

Monday, 09 December 2019

Mon 18:00 - Earth Engine Apps

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

David Gibson, Google Earth Engine

Slides

Mon 18:30 - What's New in Google Earth?

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Christiaan Adams, Google Earth

Slides

Mon 19:00 - Animations in Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Michael Dewitt, Google Earth Engine

Slides

Mon 19:30 - Flood Forecasting at Scale

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Vladimir Anisimov, Google Research

Abstract: Floods are among the most common and most deadly natural disasters in the world. The vast majority of flood-related casualties occur in developing countries, where providing spatially accurate forecasts is a challenge due to scarcity of data and lack of funding. I will describe Google's flood warning system, which provides flood extent forecast maps covering several flood-prone regions in India, and aims to eventually expand globally.

Tuesday, 10 December 2019

Tue 10:00 - Detecting Changes in Circumpolar Lake Color with Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Catherine Kuhn, University of Washington

Abstract: Over a million lakes dot the high northern latitudes, providing important wildlife habitat and supporting aquatic food webs. As high northern latitudes continue to warm at twice the global rate, the fate of water quality in these lakes is uncertain. This research documents pan-arctic changes in arctic and boreal lake color using the Collection 1 Landsat archive and Google Earth Engine.

Tue 10:30 - Raycing on the Sun

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Jessi Beck, Google Earth

Tue 11:00 - Water Clarity, Gold Mining, and Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

David Lutz, Dartmouth College

Abstract: Google Earth Engine (GEE) provides a robust framework for analyzing water clarity and understand changing environmental conditions. In the Peruvian Amazon, we examined changes in suspended sediment (SSC) in tropical rivers from artisanal gold mining. In the latter case, we found a ten-fold increase in SSC which is likely having drastic impacts on aquatic biodiversity.

Slides

Tue 11:30 - A Drought Reporting Tool on the Navajo Nation

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Amber Jean McCullum, NASA Ames Research Center

Abstract: On the Navajo Nation (NN) water is scarce, highly variable, and drought declarations are common. This project harnessed the power ClimateEngine.org to create the Drought Severity Evaluation Tool (DSET), for improved drought reporting that integrates remotely-sensed, modeled, and in-situ data on the NN. Tool co-development was made possible through an iterative and engaged knowledge sharing process between the Navajo Nation Department of Water Resources, NASA’s Applied Sciences Program, and the Desert Research Institute.

Slides

Tue 12:00 - Mapping Irrigation with TPU’s, Google Earth Engine, and Colab

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Jeremy Rapp, Michigan State University

Abstract: Irrigation is the major consumptive user of freshwater globally. Mapping the presence of irrigation using high resolution imagery is a necessary step in understanding how human driven land use change impacts planetary scale processes. We developed an artificial neural network (ANN) classification pipeline leveraging remotely sensed imagery, environmental reanalysis products, and geophysical soil attributes. The ANN is a GEE, Google Colab, and powerful TPU backed Tensorflow driven pixel by pixel multilayer perceptron.

Tue 13:00 - A Nation-wide Wetland Inventory of Canada using Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Bahram Salehi, State University of New York, College of Environmental Science and Forestry (SUNY ESF)

Abstract: We will present our Canada-wide high resolution wetland inventory map of five major wetland classes and the method we developed based on an object-based random forest classification framework utilized in Google Earth Engine cloud computing platform. The input data is some 65000 multi-year optical and SAR images, acquired by Sentinel-1 and Sentinel-2 satellites, and field samples across the country. Results show that 19% of Canada’s land areas is covered by wetlands, most of which are peatlands.

Tue 13:30 - Streetcar 2 Subduction

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Jamie Kirkpatrick, McGill University

Abstract: Streetcar 2 Subduction is a collection of geological field trips that take users to some of the world-class geological sites of the San Francisco Bay Area. We used the creation tools in Google Earth to develop the trips to be mobile-friendly for participants who access via the Google Earth app. In this presentation, we explore the inspiration for the guide and demonstrate the advantages of Google Earth for users and developers.

Tue 14:00 - Mapping and Monitoring Floods to Build Resilience

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Beth Tellman, Chief Scientist and Co-founder, Cloud to Street

Abstract: Flood risk increase as the climate changes, land use change transforms watersheds, and populations migrate into floodplains. Satellites can track these changes over past decades and in near-real time. Cloud to Street uses Google Earth Engine to map floods from 9+ satellite sensors for governments and insurers to deliver targeted response and invest in mitigation. Come hear how flood maps have been used for refugee relocation, humanitarian aid, and flood model calibration around the world.

Slides

Tue 14:30 - Land cover change through Google Earth Engine's lens

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Robert E Kennedy, Oregon State University

Abstract: Change in land surface condition affects key ecosystem processes and is forced by diverse anthropogenic and natural drivers. To characterize land surface change, we used Google Earth Engine to perform time-series segmentation on the Landsat archive, and created yearly maps of aboveground forest biomass, land cover class, percent tree cover, and percent impervious surface at 30m spatial resolution for the conterminous United States (CONUS).

Tue 15:00 - Advanced Drought and Natural Resource Monitoring with ClimateEngine.org

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Justin Huntington, Desert Research Institute

Abstract: Climate Engine is web application that programmers and non-programmers alike can use to easily process satellite and climate data in combination, and create maps and time series viewable in a web browser and available for download using Google Earth Engine. This presentation will highlight how Climate Engine it is currently being used by both public and private sectors for drought monitoring, early warning, and natural resource management.

Tue 15:30 - 48 years of glacier change in Alaska and the Yukon

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Mark Fahnestock, Geophysical Institute at the University of Alaska Fairbanks

Abstract: Nearly a half century of glacier flow and change can be visualized using Landsat data compiled into annual mosaics using Google Earth Engine. The resulting record of large-scale change, climate-induced retreat, and active rearrangement of ice flow provides our first comprehensive look at this part of Earth’s evolving ice cover.

Tue 16:00 - Raining Data from the Cloud: Using Google Earth Engine for Flood Monitoring

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Amanda Markert, NASA SERVIR Science Coordination Office, University of Alabama in Huntsville

Abstract: Remotely sensed data serves as a critical resource for generating flood maps used in disaster efforts to evaluate flood extent in areas where information and access is limited, however, managing data presents a barrier to monitoring events in a timely manner. This talk overviews SERVIR-Mekong’s development of a flood service for Myanmar leveraging Google Earth Engine to generate multi-sensor, sub-weekly flood maps enabling users to switch their focus from handling data to disaster response.

Slides

Tue 16:30 - A cloud free Landsat data set in Google Earth Engine at continental-scale: development of new high resolution land products

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Álvaro Moreno Martínez, Image Processing Laboratory (IPL), Universitat de València, Spain, Numerical Terradynamic Simulation Group (NTSG), University of Montana, Missoula, USA

Abstract: A new gap filling method has been optimized to take advantage of Google Earth Engine (GEE) and provides accurate gap free Landsat reflectance estimates and their associated uncertainties. To illustrate the potential of this reflectance data set, we trained machine learning models to produce remote sensing-based land information. These new land products could endorse an enhanced understanding of biodiversity and the Earth system with a high level of spatial detail and no data gaps.

Slides

Tue 17:00 - Mapbiomas: mapping land-use and land-cover dynamics in Brazil with big remote sensing data and artificial intelligence

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Laerte Ferreira & Leandro Parente, Image Processing and GIS Lab - LAPIG / UFG

Abstract: In spite of the significant progress Brazil has made regarding the monitoring of land-cover changes in the Amazon, existing countrywide efforts to map land use and land cover (LULC) changes are both temporally and spatially sparse. In order to overcome this gap of systematic annual LULC information at a national level, a multi-disciplinary network, called MapBiomas, was established, with the goal of reconstructing LULC information for the entire brazilian territory on an annual basis (from 1985), based on random forest applied to the Landsat Data Archive using Google Earth Engine. With an average overall accuracy of circa 90%, our maps indicate that between 1985 and 2017 Brazil lost about 71 Mha of natural vegetation, mostly to cattle ranching and agriculture activities, while another 86 Mha of the converted native vegetation is undergoing some level of regrowth. In parallel with MapBiomas' continued effort to produce increasingly accurate and current map collections, we developed the NextGenMap project, focused on establishing the next generation of land use and land cover mapping and monitoring tools, using high-res images (i.e. Planet data), cloud-computing platforms (i.e. Google Earth Engine and Google Cloud Computing) and machine learning and deep-learning techniques (e.g. Random Forest, U-Net and LSTM). our results suggest that deep learning approaches can be used to map large geographic regions, considering a wide variety of satellite data (e.g., PlanetScope, Sentinel-2, Landsat-8).

Slides

Wednesday, 11 December 2019

Wed 10:00 - Earth Engine Improv

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Tyler Erickson Google Earth Outreach

Wed 10:30 - Mapping surface water continuously across optical image collections using Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Christopher Soulard, U.S. Geological Survey, Western Geographic Science Center

Wed 11:00 - Mapping annual urban and phenology dynamics on Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Yuyu Zhou, Iowa State University

Abstract: Changes in urban environments and ecosystems play important roles in sustainable urban development. Landsat observations, together with the planetary-scale platform, provide the possibility to monitor these changes across large areas. We developed new algorithms to characterize dynamics of urban extent and phenology and implemented them on GEE for the US over the last three decades. The resulted products offer new datasets for relevant urban studies such as modeling urbanization and investigating ecosystem responses to urbanization.

Wed 11:30 - 40+ years of agriculture in the Andes’ Rain shadow

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Zachary Brecheisen, postdoctoral researcher at Purdue University

Abstract: Leveraging long-term accumulated Landsat remote sensing, Google Earth Engine, declassified historic satellite imagery, and open source statistical software we are working to recapitulate landcover change and environmental history across the Arequipa Region of southern Peru. As part of the Arequipa food, water, energy, and environment Nexus this landcover mapping work is helping us to understand the dynamic landscape context though NLCD analogue maps and chronosequence construction of landcover change.

Wed 12:00 - OpenET: Filling the Biggest Data Gap in Water Management

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Forrest Melton, NASA & Justin Huntington, Desert Research Institute

Abstract: Maximizing the benefits of our water supplies requires careful measurement of their availability and use. For irrigated agriculture, evapotranspiration (ET) is a measure of the water used to grow food, and is the biggest water use in the western US and most arid environments around the world. Accurate, transparent and easily accessible information on ET has the potential to change the way we measure and manage water in the western US. The OpenET project is developing a shared platform built on EarthEngine for data processing, visualization and distribution to provide automated and widely accessible ET data at user-defined scales and timeframes. We present an overview of the OpenET framework, user interface and application programming interface, and describe the user-driven design approach employed by the project.

Slides

Wed 12:30 - Making beautiful maps with QGIS and Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Gennadii Donchyts, Deltares

Abstract: Google Earth Engine (EE) is a great cloud-based tool for planetary-scale geospatial data processing and QGIS is one of the best desktop geoprocessing systems available, used by hundreds of thousands of GIS professionals and enthusiasts. This talk will present a preview version of the EE QGIS plugin that allows running EE scripts directly within the QGIS environment using Python API, seamlessly combining local and remote geospatial datasets to create beautiful maps.

Wed 13:00 - Visualizing urban air quality measurements

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Logan Mitchell, Department of Atmospheric Sciences, University of Utah

Abstract: We’re using Google Street View cars and public transit light rail trains to monitor air quality and greenhouse gases in Salt Lake City, Utah and visualizing the data in Google Earth. This presentation will talk about case studies that have been used to engage stakeholders and policymakers, as well as where the analysis is headed: hyperlocal source apportionment to identify local pollution sources.

Wed 13:30 - Observing marine dead zones from space

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Yingjie Li, Center for Systems Integration and Sustainability, Michigan State University.

Abstract: At least 500 dead zones have been reported across global coasts, threatening biodiversity, fisheries and billions of people’s livelihoods. However, knowledge about how global dead zones changed over time and space is still lacking. We aim to fill the gap by integrating field observations and satellite imagery in GEE to map the dynamics of global dead zone over the past two decades. This work will assist global coastal assessments and inform conservation for achieving SDGs.

Wed 14:00 - World Settlement Footprint – a novel suite supporting the analysis of global urbanization

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Mattia Marconcini, DLR, Earth Observation Center

Abstract: Reliably monitoring global urbanization is of key importance for supporting a variety of applications involving the analysis of human presence. To this purpose, we recently launched the World Settlement Footprint (WSF), i.e. a novel suite of high spatial resolution global products derived from open-and-free Big Earth Data aimed at outlining current and past extent of human settlements, as well as characterizing some of their key features like percent impervious surface, building height, and population density.

Wed 14:30 - NOAA Data and Google: Think Big and Think Open

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Chase Long and Jonathan O'Neil, NOAA

Abstract: NOAA collects more data than almost any other federal agency, and NOAA’s mission is largely focused on making those data available to the public in ways that are useful, actionable, and dependable. As demand for access to NOAA’s valuable data sets increases, innovative partnerships such as the Big Data Project (BDP) demonstrate new ways to make these data available to users around the world. Google has partnered with NOAA through the BDP to provide high-demand data sets such as GOES satellite data and NEXRAD radar with deep archive availability and low latency for current observations. These data are being used by the public, by researchers, and by innovative entrepreneurs in new ways because of the power of Google Cloud and tools. The partnership between NOAA and Google demonstrates one way that NOAA is fulfilling its mission to provide the public with broad access to high quality data at a time when the volume of that data is growing rapidly and budgets are remaining flat at best. Google’s technology and know-how have helped to serve NOAA data to more people than ever before, whilst enabling new businesses to leverage AI and cloud tools to use those data to meet the needs of a wide array of customers. Partnerships like the Big Data Project show just how much can be achieved when the public and private sectors work together.

Wed 15:00 - Women in Geospatial, Ladies of Landsat, SERVIR - Strong and inclusive communities in geosciences

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Bente Lilja Bye, BLB (Women in Geospatial) Andréa Puzzi Nicolau, UAH (Ladies of Landsat) Emily Adams, UAH (SERVIR)

Wed 15:30 - Pangeo and CMIP6: Big data climate science in the cloud

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Joe Hamman, Pangeo

Abstract: Climate scientists are awaiting the data from the sixth iteration of the Coupled Model Intercomparison Project (CMIP6) which represent the most detailed predictions ever made about our planet’s future. Despite the social and scientific value of this data, its sheer size (~20 petabytes) and complexities will present significant challenges to many users. We highlight an ongoing collaboration between Google and the Pangeo Project to provide the CMIP6 dataset as part of Google’s Public Dataset Program as well as computational tools for efficiently working with the data.

Wed 16:00 - What's new in Google Earth?

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Andrew Chang, Google Earth

Wed 16:30 - Modeling wildfire dynamics using FLAM coupled with deep learning methods on the Google Earth Engine platform

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Florian Hofhansl, Ecosystems Services and Management (ESM) Program, International Institute for Applied Systems Analysis (IIASA)

Abstract: We improve the accuracy of modeling burned areas using the FLAM model by identifying the hidden relationships between human and natural impacts on wildfire suppression efficiency applying the deep learning-based methods. FLAM uses a process-based fire parameterization algorithm taking into account daily weather (www.iiasa.ac.at/flam). The coupled approach is implemented using the Google Earth Engine platform that provides flexibility in terms of input data sets and visualization tools. We will present a case study for Indonesia.

Slides

Wed 17:00 - Using Google Resources for Mapping Air Emissions in Indian Cities

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Sarath Guttikunda Urban Emissions

Thursday, 12 December 2019

Thu 10:00 - What's new in Google Earth Engine?

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Matt Hancher, Google Earth Engine

Thu 10:30 - Characterizing air pollution patterns in Houston, Texas

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

David J. Miller, Environmental Defense Fund

Abstract: Diverse urban air pollution sources contribute to spatially variable atmospheric concentrations, with public health implications. We analyzed nine months of repeated on-road mobile measurements of black carbon, particle number and nitrogen oxides within 24 census tracts across Houston, Texas. We identified and quantified elevated pollutant concentrations 2x to >9x the sampling domain median, including nearby metal recycling and concrete batch plant facilities. These results have implications for using mobile monitoring to develop source mitigation policies.

Thu 11:00 - Monitoring surface water and energy fluxes in Brazil

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Anderson Ruhoff, Universidade Federal do Rio Grande do Sul

Abstract: Due to increasing rates of deforestation and cropland expansion in Brazil, we investigated the long-term effects of land use and land cover changes on surface fluxes. We combined land surface algorithms, remote sensing and global meteorological data using Google Earth Engine cloud computing. Our results demonstrate a potential of remote sensing models to estimate surface fluxes at regional scale to assess changes in hydrometeorological processes associated to land cover changes.

Thu 11:30 - Using GEE and MODIS ET to detect watershed disturbance

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

James Coll, University of Kansas, Department of Geography and Atmospheric Science, Mike Johnson University of California Santa Barbara, Department of Geography

Abstract: In models where ET is a function of land cover such as the North American Land Data Assimilation System or the National Water Model, it is important to identify where land cover has changed. By capitalizing on Google Earth Engine’s ability to accelerate the time to science, we develop ET signatures for all California watersheds using MOD16 and search for how deviations can be used to identify time and types of disturbance.

Slides

Thu 12:00 - Google Street View and the health burden of air pollution

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Veronica Southerland, MPH, George Washington University, Milken Institute School of Public Health & Ananya Roy, ScD Environmental Defense Fund

Abstract: Estimating air pollution health impacts at the hyper-local scale is now possible with pollutant concentrations derived from satellite remote sensing and mobile monitoring. The Environmental Defense Fund and Google Earth Outreach conducted mobile monitoring in the Bay Area using Google Street View (GSV) cars. We estimate health impacts in the Bay Area using concentrations measured by GSV cars. Hyper-local risk estimates may help local decision-makers understand how air pollution affects different neighborhoods and populations.

Slides

Thu 12:30 - Earth Engine, Earth Engine, Earth Engine...

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Michael Dewitt, Google Earth Engine

Thu 13:00 - Using Google Earth Engine to Delineate Watersheds

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Scott Haag^1,2 Ali Shokoufandeh¹, Kiran McCulloch^{1 1}Drexel University ²Children’s Hospital of Philadelphia

Abstract: Google Earth Engine (GEE) is a cloud-based image repository and tile processing environment. Existing algorithms to delineate watersheds require path searches across Digital Elevation Models. These algorithms will not work in GEE because they require information to be exchanged across image tiles. In this presentation we demonstrate a novel data model (the Modified Nested Set model) that can be used to overcome the intra-tile information transfer issue, enabling high-performance delineation of watersheds in GEE.

Slides

Thu 13:30 - Long-term changes at multiple scales with the BULC algorithm

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Jeffrey Cardille, McGill University

Abstract: How well can images from the earliest years of the satellite record inform a time series at a national scale? Can the limitations of MSS imagery be overcome to produce plausible maps of land cover, disturbed areas, forest recovery, and forest stability beginning in 1972, when Landsat 1 was first launched? Here we look back with a survey of opportunities and constraints for that low-information era, and summarize some of our other exciting EE work!

Thu 14:00 - Dynamic World Project - Automating the Development of a High Resolution Land Use, Land Cover Dataset for the Globe.

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Joe Mazzariello, National Geographic Society

Abstract: In its efforts to contribute to protecting 30% of the planet by 2030, National Geographic Society has partnered with Google, and WRI on the Dynamic World Project. This project utilizes a unique methodology and a machine learning/software approach to land cover mapping based on a training set of 1 billion pixels in densely labeled tiles. NGS labs has coordinated the team needed to create the first land use / land cover training dataset on ESA Sentinel-2 10m/pixel data available through Google Earth Engine.

Slides

Thu 14:30 - Leveraging Google Earth Engine for rangelands monitoring

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Bo Zhou, University of California, Los Angeles

Abstract: Mapping and monitoring of soil cover, vegetation structure, and various native and non-native species is critical for rangeland management. With the availability of Google Earth Engine (GEE), the capability now exists to conduct planetary-scale analysis, including mapping of rangeland indicators. Combined with the collection of large amounts of in situ data in the western U.S., GEE can enable the prediction of surface conditions at times and places no in situ data are available

Thu 15:00 - Finding Antarctica's Fractures

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Ching-Yao Lai , Lamont-Doherty Earth Observatory, Columbia University

Abstract: Atmospheric warming threatens to accelerate the retreat of the Antarctic Ice Sheet by increasing surface melting and facilitating ‘hydrofracturing’, where meltwater flows into and enlarges fractures, triggering ice-shelf collapse. Through a deep neural network we obtain the first map of fracture locations on all Antarctic ice shelves, and demonstrate close agreement between the distribution of fracture and our physics-based model. Our result can be useful for including observationally-constrained ice damage in ice-sheet models and our methodology can also be adopted to monitor growth of fractures.

Thu 15:30 - Deforestation drivers in the Amazon using SAR

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Andrea Nicolau, University of Alabama in Huntsville / NASA SERVIR Science Coordination Office

Abstract: The Amazon is the largest tropical rainforest globally and deforestation resulting from land changes poses major concern for sustainable resource management. With the power of Google Earth Engine, other open-source platforms, and freely-available radar time-series data a classification decision tree for the monitoring of land use and cover dynamics in the Peruvian Amazon was created. These results can aid current monitoring systems and the identification of deforestation drivers by leveraging radar technology in open-source platforms.

Slides

Thu 16:00 - Enabling Annual Land Cover Mapping in the Eastern and Southern Africa Region

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Phoebe Oduor,

Abstract: Land cover mapping is key to a lot of development decisions. The frequency with which this is done enables governments to make better decisions. In the recent past, there have been a lot of changes with how land cover mapping is done. The increase in accessible data enabled time series mapping. Cloud resources, such as google, have enabled rapid processing. We shall demonstrate how we applied google earth resources in improving land cover mapping processes.

Thu 16:30 - Mapping Winter Cover Crop Performance in Maryland

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Alison Thieme, NASA DEVELOP (SSAI) and University of Maryland, College Park, and Sean McCartney, NASA DEVELOP (SSAI)

Abstract: Farmers in Maryland are incentivized to plant winter cover crops to retain nutrients in fields, reduce runoff, and improve soil structure. These positive impacts are directly related to crop performance affected by environmental and agronomic variables. Google Earth Engine allowed us to analyze Landsat 5/8 and Sentinel-2 imagery of fields over to evaluate biomass and percent cover associated with different agronomic management strategies. This informs Maryland Department of Agriculture’s adaptive management in the Chesapeake Bay.

Thu 17:00 - What's new in Google Earth?

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Andy Blank, Google Earth

Friday, 13 December 2019

Fri 10:00 - Using GEE for Decision Making in developing countries in East Africa

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

James Nyaga Wanjohi, Regional Centre for Mapping of Resources for Development (RCMRD)

Abstract: SERVIR is a USAID-NASA funded program that helps developing countries use satellite data to address critical challenges in food security, water resources, weather and climate, land use and natural hazards. At RCMRD, where SERVIR-ESA is hosted, we have leveraged on the capabilities of GEE to create products and solutions that have been used for decision making to address some of these critical challenges in our region of East Africa.

Fri 10:30 - Wheat yield prediction at a county and field scale with deep learning, machine learning, and Google Earth Engine

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Juan Cao, Beijing Normal University

Abstract: We processed all data on Google Earth Engine and then used three typical deep learning models and a machine learning model to predict wheat yields over large areas. Overall, all four models performed well at a county level and field level. Our findings provided a new scalable, simple and inexpensive framework for estimating crop yield on a regional scale based on Google Earth Engine platform.

Fri 11:00 - Global Framework for Air Quality Monitoring

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Nathan Pavlovic, Sonoma Technology, Inc

Abstract: Globally, nearly 4.5 million premature deaths per year can be attributed to fine particulate matter (PM2.5) pollution. Satellite observations offer the potential to provide a comprehensive view of PM2.5 in near-real time when combined with data from ground-level monitors and air quality models. In this presentation, we will share the results of fusing satellite data with surface observations to estimate ground-level air pollution using Google Earth Engine. These results show promise for filling the data gap in Africa.

Fri 11:30 - Environmental Observatory to Support Eco-Entrepreneurship in Benin

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Ufuoma Ovienmhada and Danielle Wood, Space Enabled Research Group, MIT Media Lab

Abstract: Coastal ecosystems provide many benefits to humans in the form of transportation, subsistence, and economic opportunity. This project develops techniques that enable low cost, data-informed governance of human-environmental coastal ecosystems using earth observation and Google Earth Engine. The talk applies these methods to a case study in Benin Republic, a country that has been harmed by the infestation of an invasive plant species known as the water hyacinth.

Fri 12:00 - 30-meter Leaf Area Index Mapping for the Contiguous US

Moscone North, Exhibit Hall | Google Exhibit Booth Presentation

Yanghui Kang, University of Wisconsin-Madison, Department of Geography

Abstract: Leaf Area Index (LAI) is an essential ecological variable characterizing the amount of foliage in vegetation. Satellite images have been used to derived Long-term LAI products for ecosystem monitoring and modeling, but the coarse spatial resolution (0.5km – 4km) is inadequate for many applications. This talk will present a scalable approach to map LAI at 30-meter resolution for the Contiguous US, leveraging Landsat images, MODIS LAI, and machine learning algorithms on Google Earth Engine.

Slides