Conference Agenda

Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

Session Overview
211RA: Advances in land monitoring for sustainable development - Part A
Wednesday, 24/Apr/2019:
11:15am - 12:45pm

Session Chair: Patrick Hostert
Session Chair: Veronique De Sy
Location: MB-220
Main Building, room 220, second floor, west wing, 154 (+22) seats
Session Topics:
What do people want from land?

Session Abstract

This session will explore the state-of-the-art in land monitoring, focussing on the use of freely available, global remote sensing data which can underpin land use assessments and contribute to solving pressing questions in the land use sector. The Paris Climate Agreement recognizes the importance of reducing emissions from deforestation and forest degradation, through the REDD+ results-based payments mechanism. Monitoring of forest cover and forest parameters (such as biomass) is an essential component. At the same time, increasing agricultural production is key in reaching the zero hunger SDG target, and, as well as increasing production (on existing agricultural land), area of land under agriculture is also expanding. Some monitoring needs related to agricultural land area and forest area can be potentially addressed by land cover change maps, which can provide information on for example conversions from forest to agriculture. This constantly evolving field (in terms of stakeholder demands, and available data from upcoming space missions and technology) calls for an expert-driven platform where researchers and other stakeholders can share knowledge, and work together to develop best-practices for land monitoring. GOFC-GOLD (Global Observation for Forest Cover and Land Dynamics) is currently supporting advances in the field and is developing guidance to overcome challenges including improving monitoring using interpretation of high-resolution satellite data for land cover (change) mapping and validation. Accurate, transparent and reproducible methods and results are required for monitoring purposes, and to enable decision makers (including forest managers, private sector, civil society and government agencies) to identify, plan and develop priority interventions and policies. This can also identify opportunities to reduce trade-offs from competing land-uses and to increase synergies, to achieve climate-smart land use.

External Resource: - SESSION RECORDING -
Show help for 'Increase or decrease the abstract text size'
Full talk
ID: 867 / 211RA: 1
211R Advances in land monitoring for sustainable development
Keywords: Land-use change, Land-cover detection, Pseudo ground-truth, Land-use modelling

Using pseudo ground-truth data for constructing historical land-cover maps: An Australian case study

Marco Rodrigo Calderon Loor, Brett Bryan, Michalis Hadjikakou

Deakin University, Australia

Global land-use dynamics are influenced by variability in biophysical and socioeconomic factors. Increase in urban sprawl alongside Australian population is expected to pose a threat over Australian food production and biodiversity. Modelling of these impacts requires accurate land-cover/land-use maps at different time slices. However, historically consistent maps at moderate resolution are scarce. Furthermore, past ground-truth data is not always available, adding complexity to the development of land-cover maps.

We present a methodology for building historical land-cover maps using pseudo ground-truth data. Using all the Landsat historical imagery available in the Google Earth Engine (GEE) platform we developed a timeseries of seven land-cover maps for the period 1985-2015 at 30 m resolution for all Australia. Random stratified sampling was used to generate pseudo ground-truth data, with the National Dynamic Land Cover Dataset of Australia 2.1 (DLCD) acting as a reference. An iterative clustering algorithm was implemented to correct mislabelled pixels. The remaining pixels, around 30000, were used for producing the historical maps using a random forest algorithm in GEE. A further post-processing process was implemented to reduce the uncertainty and inconsistencies in the final outputs. Both the methodology and the resulting maps offer valuable insights in the land-cover/land-use change detection study as the process can be replicated at different scales.

This work is the first part of a research project aimed at improving the prediction accuracy and resolution of land-use change impacts on agriculture and biodiversity in Australia. The generated maps will be used to develop a land-use simulation model that can forecast future land uses under different scenarios for Australia. Reducing uncertainty in prediction will help us to anticipate and design strategies to prevent possible impacts on land systems. Additionally, our results will be coupled with computable general equilibrium (CGE) models and species distribution models (SDM) for predicting the impacts of trade on biodiversity.

Full talk
ID: 745 / 211RA: 2
211R Advances in land monitoring for sustainable development
Keywords: remote sensing, greenhouse gas emissions, land monitoring, SDG, climate change mitigation

Innovating land monitoring in support of climate change mitigation and sustainable land use strategies

Veronique De Sy1, Martin Herold1, Sarah Carter1, Astrid Bos1, Patric Brandt1, Nancy Harris2, Renske Hijbeek3, Marloes van Loon3, Johannes Reiche1

1Laboratory of Geo-information Science and Remote Sensing, Wageningen University, the Netherlands; 2World Resource Institute, Washington DC, USA; 3Plant Production Systems, Wageningen University, the Netherlands

The 2015 Paris Agreement requires all countries to put forward nationally determined contributions (NDCs) to mitigate climate change. Many countries have included agriculture, forestry and other land-use (AFOLU) targets in their NDCs. In addition to climate change mitigation goals, developing countries will want to reach other sustainable development goals (SDGs) such as improved food security. Countries will need AFOLU information to improve their GHG inventories, inform policy design and planning, support policy implementation and enforcement, and to assess policy performance. This is challenging in many developing countries where monitoring capacities are low, and the potential of land-based mitigation is high.

A considerable amount of publicly available, comprehensive, spatial information on land use (change) and associated carbon stocks and flows has become available. This information can greatly assist with addressing some of the monitoring challenges in the AFOLU sector. The objective of this presentation is to highlight ongoing research on the use of available spatial data and methods to support policy design and planning by looking at data-driven assessments on agriculture-forest dynamics and resulting GHG emissions. We show how different scenarios of agricultural intensification versus expansion of agriculture affect food security and total greenhouse gas emissions. Near-real time monitoring systems can assist with implementation and enforcement of zero-deforestation pledges and sustainable supply chains. In addition, we discuss how available forest cover data and tools can aid in assessing (REDD+) policy performance at the local scale, and what important considerations are regarding accuracy, uncertainty and complementarity of different datasets. Finally, we present spatially explicit greenhouse gas emissions for the tropics developed with a data-driven approach. This presentation gives more insight into climate change mitigation opportunities and possible trade-offs with broader development goals and the SDGs using data-driven approaches.

Flash talk
ID: 786 / 211RA: 3
211R Advances in land monitoring for sustainable development
Keywords: silage bags, cattle production, livestock system, remote sensing

Mapping silage bags - the missing link between agriculture and livestock

Philipp Gaertner1, Pedro Fernandez1, Ignacio Gasparri1,2

1Instituto de Ecología Regional, Tucumán, Argentina, Argentine Republic; 2Instituto de Investigación Animal del Chaco Semiárido., Tucumán, Argentina

Livestock production is the most widespread use on earth, however consequences on earth system of livestock remain misunderstood principally due to scarcity of data. In cattle production, the lack of information of non-marketable forage constitutes one of the biggest gaps.

We argue, that we can reduce this knowledge gap with the mapping of silage bag usage. Silage bag technology (used since 20 years) is based on anaerobic conservation, low in costs and easily implemented. At least 50 countries use this technology including larger producers of beef and grains (e.g. Argentina, Brazil, United State). Producers mainly using it for two purposes: dry oleaginous and cereal temporal storage, and humid storage of forage (commonly grains as maize or harvested pastures) for cattle supplementation.

In cattle production silage storage constitute a strategy for sustaining nutritional requirements and improving productivity in different situations, particularly in climates with strong seasonality.

We use high resolution satellite imagery from Sentinel-2 (06/2015 until 10/2018) and detect every silage bag within the province of Salta and Jujuy, two provinces of Northwest of Argentina (208.707 km2) with a classification accuracy of 94%. Furthermore we analyse the dynamics of silage bag usage and provide information (m3) on occurence change intensity, seasonality and recurrence.

We believe that our provided silage bag usage data opens new opportunities to better understand how the links between agriculture and cattle production could facilitate the activations of new frontiers, supporting cattle ranching in restricted conditions especially in sectors with strong aridity gradients. For more, this could be conceptualized as the manifestation of telecoupling between agriculture and cattle necessary to explain the cattle production in marginal areas. Finally, we believe that information about silos (during dry season) of silage bags could be a new metric of land use intensity for regional analyses in livestock system of subtropical climates.

Flash talk
ID: 437 / 211RA: 4
211R Advances in land monitoring for sustainable development
Keywords: land-use classification, crop types, Sentinel 2, time series, dynamic composite periods

Introducing dynamic composite periods for pixel-based crop type mapping using Sentinel-2A imagery

Daniel Doktor, Maximilian Lange, Sebastian Preidl

Helmholtz Centre for Environmental Research - UFZ, Germany

Here, we present a novel approach for a detailed land-use classification (LUC) exploiting high spatio-temporal resolution Sentinel 2 data. This approach is i) transferable between study areas and also ii) scalable from the landscape to a nationwide level. Even with short satellite revisit times LUC is hampered by frequent cloud cover and potentially varying number of satellite imagery within a study area due to the satellite's flight path. Furthermore, the spatial distribution of present land-use types may change within a study area and almost certainly between neighbouring study areas. Additionally, the spatial representation of these land-use types in available ground truth data (used for training) is not identical with the actual representation.

The presented methodology addresses the above mentioned challenges via the generation of Dynamic Composite Periods (DCP). These may vary in length and temporal spacing since they are purely data-driven based on satellite data availability and the spatial distribution of the training data. The presented framework also allows mapping a pixel-wise classification uncertainty based on respective machine learning models. We demonstrate the method's capability for the area of Germany where 17 crop types were mapped in six bio-geographical regions for the year 2016. The study could draw on extensive ground validation data as gathered by the federal states of Germany.

We identified a varying DCP number between the bio-geographical areas with different lengths and temporal spacings emphasising the need for a flexible classification procedure accommodating the respective regional situation. The growth stages of classified crop types could well be captured resulting in a high overall classification accuracy of > 80%, > 90 % for the major crop types. The classification uncertainty varied considerably between and also within the bio-geographical regions dependent on the spatial distribution of ground truth data, cloud coverage and the satellite's flight path.

Full talk
ID: 825 / 211RA: 5
211R Advances in land monitoring for sustainable development
Keywords: Land Cover, Global, COnsistency, Change, Climate, ESA CCI, C3S

Annual global land cover maps from the 1990s to 2016 for climate modelling: achievements of the Land Cover component of the ESA Climate Change Initiative evolving to the Copernicus Climate Change operational service

P. Defourny1, C. Brockmann3, C. Lamarche1, S. Bontemps1, F. Achard2, M. Boettcher3, T. De Maet1, P. Gamba4, S. Hagemann5, A. Hartley6, L. Hoffman7, G. Georgievski5, G. Kirches3, N. MacBean8, P. Mayaux2, I. Moreau1, C. Ottle8, C. Pathe9, P. Peylin7, J. Radoux1, F. Ramoino10, M. Santoro11, C. Schmullius9, E. Van Bogaert1, U. Wegmüller11, M. Zuehlke3, O. Arino10

1UCLouvain-Geomatics (Belgium), Belgium; 2Joint Research Center; 3Brockmann Consult GmbH, Germany; 4University of Pavia, Italy;; 5Max Plank Institute for Meteorology, Germany; 6MET Office, United Kingdom; 7LIST, Luxembourg; 8LSCE, France; 9University of Jena, Germany; 10ESA-ESRIN, Italy; 11GAMMA Remote Sensing, Switzerland

Essential Climate Variables were listed by the Global Climate Observing System as critical information to further understand the climate system and support climate modelling. The European Space Agency launched its Climate Change Initiative (CCI) to provide an adequate response to the set of requirements for long-term satellite-based products for climate.

Within this program, the Land Cover (LC) project aimed at revisiting all algorithms required for the generation of global LC products that are stable and consistent over time. To this end, the LC concept was revisited to deliver a set of consistent 300m global annual LC maps from the 1990s to 2015 describing the land surface into 22 classes using the FAO LC Classification System.

The entire MERIS Full and Reduced Resolution (FR and RR) archive from 2003 to 2012 was first classified into a unique 10-year baseline LC map, which is then back- and up-dated using (i) AVHRR time series from 1992 to 1999, (ii) SPOT-VGT time series from 1998 to 2012 and (iii) PROBA-V time series from 2013 to 2015. This method avoids independent classification from year to year, thus ensuring temporal and spatial consistency between successive maps. The CCI LC project also delivered climatological 7-day time series representing seasonal burned areas, vegetation and snow dynamics of the land surface, 7-day surface reflectance time series for the whole archive of MERIS FR and RR and a global map of open permanent water bodies at 150 m derived from Envisat ASAR images and auxiliary data. All products are delivered along with an aggregation tool, enabling re-projection and re-sampling as well as the translation from LC classes into Plant Functional Types for the different climate models.

The extension of the global annual LC maps series from 2016 to 2019 in a consistent manner will be ensured with PROBA-V within the operational European Commission Copernicus Climate Change Service. Each LC map of a given year will be released the next year, with accuracy assessed in the 3 months following the release.

Flash talk
ID: 551 / 211RA: 6
211R Advances in land monitoring for sustainable development
Keywords: land-use change, open science, collaborative science, modelled data products, spatio-temporal dataset

Mapping global land-use patterns and recent historical changes

Steffen Ehrmann1, Giuseppe Amatulli2, Lars Bernard3, Daniele Da Re4, Steffen Fritz5, Allessandro Gentile1, Marius Gilbert6, Stephan Mäs3, Felipe Melges1, Isabel Rosa7, Navin Ramankutty8, Longzhu Shen9, Katharina Schulze10, Ralf Seppelt11, Peter Verburg10, Florian Wolf1, Liangzhi You12, Carsten Meyer1

1German Centre for Integrative Biodiversity Research (iDiv), Germany; 2Yale School of Forestry & Environmental Studies; 3Technische Universität Dresden; 4Université Catholique de Louvain; 5International Institute for Applied Systems Analysis; 6Université Libre de Bruxelles; 7Bangor University; 8University of British Columbia; 9University of Cambridge; 10Vrije Universiteit Amsterdam; 11Helmholtz-Centre for Environmental Research; 12International Food Policy Research Institute

Addressing several environmental and socio-economic problems and monitoring progress towards multiple Sustainable Development Goals requires detailed global scale information on current patterns and historical dynamics in crop, livestock and forest production systems. Despite their bearing, gridded modelled data products (MDPs) of various land-use variables are rarely available as historical time-series, are not mutually consistent and suffer generally from low precision and accuracy. Mutual consistency across various MDPs is challenged in particular by a lack of standardization (e.g. incompatible datasets and different assumptions in algorithms or models). Recent advances in remote-sensing, such as longer and higher-quality time-series on land-cover and climate, recent progress in mobilizing and sharing sub-national land-use and production statistics, and technology supporting new forms of collaboration now offer opportunities for addressing various of the significant limitations.
The Land-Use Change Knowledge Integration Network (LUCKINet) was launched to develop a new generation of global, annual, mutually consistent, data-driven and quality-assured MDPs for several important land-use variables. LUCKINet aims to enable regular semi-automated updates and retrospective improvements of MDPs as better input data and modelling techniques become available. Therefore, we organize the collaborative development of MDPs in the form of modularized algorithms within a standardized open-science IT infrastructure. We envision to engage a growing network of contributors from the wider community in land-system science to successively increase the list of covered land-use variables, consolidate and improve the employed techniques and to make use of the data.
We present initial results from the first collaborative project of the LUCKINet:
- The state of our effort to develop a transparent, reproducible and community-driven IT infrastructure and - An (operational) alpha-version of a gridded time-series of several cropping, grazing, and forestry related variables with mutual statistical, spatial and temporal consistency spanning the past 2-3 decades.

Contact and Legal Notice · Contact Address:
Privacy Statement · Conference: OSM 2019
Conference Software - ConfTool Pro 2.6.131+TC+CC
© 2001 - 2020 by Dr. H. Weinreich, Hamburg, Germany