@article{KoehlerKuenzer2020,
  author    = {Koehler, Jonas and Kuenzer, Claudia},
  title     = {Forecasting spatio-temporal dynamics on the land surface using Earth Observation data — a review},
  series = {Remote Sensing},
  volume    = {12},
  journal   = {Remote Sensing},
  number    = {21},
  issn      = {2072-4292},
  doi       = {10.3390/rs12213513},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-216285},
  year      = {2020},
  abstract  = {Reliable forecasts on the impacts of global change on the land surface are vital to inform the actions of policy and decision makers to mitigate consequences and secure livelihoods. Geospatial Earth Observation (EO) data from remote sensing satellites has been collected continuously for 40 years and has the potential to facilitate the spatio-temporal forecasting of land surface dynamics. In this review we compiled 143 papers on EO-based forecasting of all aspects of the land surface published in 16 high-ranking remote sensing journals within the past decade. We analyzed the literature regarding research focus, the spatial scope of the study, the forecasting method applied, as well as the temporal and technical properties of the input data. We categorized the identified forecasting methods according to their temporal forecasting mechanism and the type of input data. Time-lagged regressions which are predominantly used for crop yield forecasting and approaches based on Markov Chains for future land use and land cover simulation are the most established methods. The use of external climate projections allows the forecasting of numerical land surface parameters up to one hundred years into the future, while auto-regressive time series modeling can account for intra-annual variances. Machine learning methods have been increasingly used in all categories and multivariate modeling that integrates multiple data sources appears to be more popular than univariate auto-regressive modeling despite the availability of continuously expanding time series data. Regardless of the method, reliable EO-based forecasting requires high-level remote sensing data products and the resulting computational demand appears to be the main reason that most forecasts are conducted only on a local scale. In the upcoming years, however, we expect this to change with further advances in the field of machine learning, the publication of new global datasets, and the further establishment of cloud computing for data processing.},
  language  = {en}
}
@article{RemelgadoSafiWegmann2020,
  author    = {Remelgado, Ruben and Safi, Kamran and Wegmann, Martin},
  title     = {From ecology to remote sensing: using animals to map land cover},
  series = {Remote Sensing in Ecology and Conservation},
  volume    = {6},
  journal   = {Remote Sensing in Ecology and Conservation},
  number    = {1},
  doi       = {10.1002/rse2.126},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-225200},
  pages     = {93-104},
  year      = {2020},
  abstract  = {Land cover is a key variable in monitoring applications and new processing technologies made deriving this information easier. Yet, classification algorithms remain dependent on samples collected on the field and field campaigns are limited by financial, infrastructural and political boundaries. Here, animal tracking data could be an asset. Looking at the land cover dependencies of animal behaviour, we can obtain land cover samples over places that are difficult to access. Following this premise, we evaluated the potential of animal movement data to map land cover. Specifically, we used 13 White Storks (Cicona cicona) individuals of the same population to map agriculture within three test regions distributed along their migratory track. The White Stork has adapted to foraging over agricultural lands, making it an ideal source of samples to map this land use. We applied a presence-absence modelling approach over a Normalized Difference Vegetation Index (NDVI) time series and validated our classifications, with high-resolution land cover information. Our results suggest White Stork movement is useful to map agriculture, however, we identified some limitations. We achieved high accuracies (F1-scores > 0.8) for two test regions, but observed poor results over one region. This can be explained by differences in land management practices. The animals preferred agriculture in every test region, but our data showed a biased distribution of training samples between irrigated and non-irrigated land. When both options occurred, the animals disregarded non-irrigated land leading to its misclassification as non-agriculture. Additionally, we found difference between the GPS observation dates and the harvest times for non-irrigated crops. Given the White Stork takes advantage of managed land to search for prey, the inactivity of these fields was the likely culprit of their underrepresentation. Including more species attracted to agriculture - with other land-use dependencies and observation times - can contribute to better results in similar applications.},
  language  = {en}
}
@article{UllmannSchmittRothetal.2014,
  author    = {Ullmann, Tobias and Schmitt, Andreas and Roth, Achim and Duffe, Jason and Dech, Stefan and Hubberten, Hans-Wolfgang and Baumhauer, Roland},
  title     = {Land Cover Characterization and Classification of Arctic Tundra Environments by Means of Polarized Synthetic Aperture X- and C-Band Radar (PolSAR) and Landsat 8 Multispectral Imagery — Richards Island, Canada},
  doi       = {10.3390/rs6098565},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-113303},
  year      = {2014},
  abstract  = {In this work the potential of polarimetric Synthetic Aperture Radar (PolSAR) data of dual-polarized TerraSAR-X (HH/VV) and quad-polarized Radarsat-2 was examined in combination with multispectral Landsat 8 data for unsupervised and supervised classification of tundra land cover types of Richards Island, Canada. The classification accuracies as well as the backscatter and reflectance characteristics were analyzed using reference data collected during three field work campaigns and include in situ data and high resolution airborne photography. The optical data offered an acceptable initial accuracy for the land cover classification. The overall accuracy was increased by the combination of PolSAR and optical data and was up to 71\% for unsupervised (Landsat 8 and TerraSAR-X) and up to 87\% for supervised classification (Landsat 8 and Radarsat-2) for five tundra land cover types. The decomposition features of the dual and quad-polarized data showed a high sensitivity for the non-vegetated substrate (dominant surface scattering) and wetland vegetation (dominant double bounce and volume scattering). These classes had high potential to be automatically detected with unsupervised classification techniques.},
  language  = {en}
}
@article{LandmannSchrammColditzetal.2010,
  author    = {Landmann, Tobias and Schramm, Matthias and Colditz, Rene R. and Dietz, Andreas and Dech, Stefan},
  title     = {Wide Area Wetland Mapping in Semi-Arid Africa Using 250-Meter MODIS Metrics and Topographic Variables},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-68628},
  year      = {2010},
  abstract  = {Wetlands in West Africa are among the most vulnerable ecosystems to climate change. West African wetlands are often freshwater transfer mechanisms from wetter climate regions to dryer areas, providing an array of ecosystem services and functions. Often wetland-specific data in Africa is only available on a per country basis or as point data. Since wetlands are challenging to map, their accuracies are not well considered in global land cover products. In this paper we describe a methodology to map wetlands using well-corrected 250-meter MODIS time-series data for the year 2002 and over a 360,000 km2 large study area in western Burkina Faso and southern Mali (West Africa). A MODIS-based spectral index table is used to map basic wetland morphology classes. The index uses the wet season near infrared (NIR) metrics as a surrogate for flooding, as a function of the dry season chlorophyll activity metrics (as NDVI). Topographic features such as sinks and streamline areas were used to mask areas where wetlands can potentially occur, and minimize spectral confusion. 30-m Landsat trajectories from the same year, over two reference sites, were used for accuracy assessment, which considered the area-proportion of each class mapped in Landsat for every MODIS cell. We were able to map a total of five wetland categories. Aerial extend of all mapped wetlands (class "Wetland") is 9,350 km2, corresponding to 4.3\% of the total study area size. The classes "No wetland"/"Wetland" could be separated with very high certainty; the overall agreement (KHAT) was 84.2\% (0.67) and 97.9\% (0.59) for the two reference sites, respectively. The methodology described herein can be employed to render wide area base line information on wetland distributions in semi-arid West Africa, as a data-scarce region. The results can provide (spatially) interoperable information feeds for inter-zonal as well as local scale water assessments.},
  subject      = {Geologie},
  language  = {en}
}