A spatiotemporal machine learning framework for automated prediction and analysis of long-term Land Use/Land Cover dynamics is presented. The framework includes: (1) harmonization and preprocessing of spatial and spatiotemporal input datasets (GLAD Landsat, NPP/VIIRS) including five million harmonized LUCAS and CORINE Land Cover-derived training samples, (2) model building based on spatial k-fold cross-validation and hyper-parameter optimization, (3) prediction of the most probable class, class probabilities and model variance of predicted probabilities per pixel, (4) LULC change analysis on time-series of produced maps. The spatiotemporal ensemble model consists of a random forest, gradient boosted tree classifier, and an artificial neural network, with a logistic regressor as meta-learner. The results show that the most important variables for mapping LULC in Europe are: seasonal aggregates of Landsat green and near-infrared bands, multiple Landsat-derived spectral indices, long-term surface water probability, and elevation. Spatial cross-validation of the model indicates consistent performance across multiple years with overall accuracy (a weighted F1-score) of 0.49, 0.63, and 0.83 when predicting 43 (level-3), 14 (level-2), and five classes (level-1). Additional experiments show that spatiotemporal models generalize better to unknown years, outperforming single-year models on known-year classification by 2.7% and unknown-year classification by 3.5%. Results of the accuracy assessment using 48,365 independent test samples shows 87% match with the validation points. Results of time-series analysis (time-series of LULC probabilities and NDVI images) suggest forest loss in large parts of Sweden, the Alps, and Scotland. Positive and negative trends in NDVI in general match the land degradation and land restoration classes, with "urbanization" showing the most negative NDVI trend. An advantage of using spatiotemporal ML is that the fitted model can be used to predict LULC in years that were not included in its training dataset, allowing generalization to past and future periods, e.g. to predict LULC for years prior to 2000 and beyond 2020. The generated LULC time-series data stack (ODSE-LULC), including the training points, is publicly available via the ODSE Viewer. Functions used to prepare data and run modeling are available via the eumap library for Python.

Department of Geodesy and Geoinformatics Faculty of Civil Engineering University of Belgrade Belgrade Serbia

Department of Geomatics Faculty of Civil Engineering Czech Technical University of Prague Prague Czech Republic

Envirometrix Wageningen The Netherlands

GILAB Belgrade Serbia

MultiOne Zagreb Croatia

National Meteorological Administration of Romania Bucharest Romania

OpenGeoHub Wageningen The Netherlands

Technical University of Civil Engineering Bucharest Bucharest Romania

Terrasigna Bucharest Romania

Zobrazit více v PubMed

Arino O, Ramos Perez JJ, Kalogirou V, Bontemps S, Defourny P, Van Bogaert E. Global land cover map for 2009 (GlobCover 2009) Paris, Louvain-la-Neuve: European Space Agency (ESA) & Université catholique de Louvain (UCL); 2012.

Batista e Silva F, Lavalle C, Koomen E. A procedure to obtain a refined European land use/cover map. Journal of Land Use Science. 2013;8(3):255–283. doi: 10.1080/1747423X.2012.667450. DOI

Bossard M, Feranec J, Otahel J. CORINE land cover technical guide: addendum 2000. Vol. 40. European Environment Agency; Copenhagen: 2000.

Breiman L. Random forests. Machine Learning. 2001;45(1):5–32. doi: 10.1023/A:1010933404324. DOI

Breidenbach J, Ellison D, Petersson H, Korhonen KT, Henttonen HM, Wallerman J, Fridman J, Gobakken T, Astrup R, Næsset E. Harvested area did not increase abruptly—how advancements in satellite-based mapping led to erroneous conclusions. Annals of Forest Science. 2022;79:2. doi: 10.1186/s13595-022-01120-4. DOI

Buchhorn M, Lesiv M, Tsendbazar N-E, Herold M, Bertels L, Smets B. Copernicus global land cover layers –Collection 2. Remote Sensing. 2020;12(6):1044. doi: 10.3390/rs12061044. DOI

Buck O, Haub C, Woditsch S, Lindemann M, Kleinwillinghöfer L, Hazeu G, Kosztra B, Kleeschulte S, Arnold S, Hölzl M. European Environment Agency Technical Report, European Environment Agency and the European Environment Information and Observation Network; Copenhagen: 2015. Analysis of the LUCAS nomenclature and proposal for adaptation of the nomenclature in view of its use by the Copernicus land monitoring services.

Buus-Hinkler J, Hansen BU, Tamstorf MP, Pedersen SB. Snow-vegetation relations in a High Arctic ecosystem: Inter-annual variability inferred from new monitoring and modeling concepts. Remote Sensing of Environment. 2006;105(3):237–247. doi: 10.1016/j.rse.2006.06.016. DOI

Buyantuyev A, Wu J. Effects of thematic resolution on landscape pattern analysis. Landscape Ecology. 2007;22(1):7–13. doi: 10.1007/s10980-006-9010-5. DOI

Caldern-Loor M, Hadjikakou M, Bryan BA. High-resolution wall-to-wall land-cover mapping and land change assessment for Australia from 1985 to 2015. Remote Sensing of Environment. 2021;252:112–148. doi: 10.1016/j.rse.2020.112148. DOI

Castilla G, Larkin K, Linke J, Hay GJ. The impact of thematic resolution on the patch-mosaic model of natural landscapes. Landscape Ecology. 2009;24(1):15–23. doi: 10.1007/s10980-008-9310-z. DOI

Ceccherini G, Duveiller G, Grassi G, Lemoine G, Avitabile V, Pilli R, Cescatti A. Abrupt increase in harvested forest area over Europe after 2015. Nature. 2020;583(7814):72–77. doi: 10.1038/s41586-020-2438-y. PubMed DOI

Chen J, Chen J, Liao A, Cao X, Chen L, Chen X, He C, Han G, Peng S, Lu M. Global land cover mapping at 30 m resolution: a POK-based operational approach. ISPRS Journal of Photogrammetry and Remote Sensing. 2015;103:7–27. doi: 10.1016/j.isprsjprs.2014.09.002. DOI

Chen T, Guestrin C. Xgboost: a scalable tree boosting system. Proceedings of the 22nd ACM sigkdd international conference on knowledge discovery and data mining; New York. 2016. pp. 785–794.

Conway T. The impact of class resolution in land use change models. Computers, Environment and Urban Systems. 2009;33(4):269–277. doi: 10.1016/j.compenvurbsys.2009.02.001. DOI

d’Andrimont R, Verhegghen A, Meroni M, Lemoine G, Strobl P, Eiselt B, Yordanov M, Martinez-Sanchez L, van der Velde M. LUCAS Copernicus 2018: earth-observation-relevant in situ data on land cover and use throughout the European Union. Earth System Science Data. 2021;13(3):1119–1133. doi: 10.5194/essd-13-1119-2021. DOI

d’Andrimont R, Yordanov M, Martinez-Sanchez L, Eiselt B, Palmieri A, Dominici P, Gallego J, Reuter HI, Joebges C, Lemoine G. Harmonised LUCAS in-situ land cover and use database for field surveys from 2006 to 2018 in the European Union. Scientific Data. 2020;7(1):1–15. doi: 10.1038/s41597-019-0340-y. PubMed DOI PMC

Defazio A, Bach F, Lacoste-Julien S. SAGA: a fast incremental gradient method with support for non-strongly convex composite objectives. Advances in Neural Information Processing Systems. 2014;27:1646–1654.

Duveiller G, Caporaso L, Abad-Vias R, Perugini L, Grassi G, Arneth A, Cescatti A. Local biophysical effects of land use and land cover change: towards an assessment tool for policy makers. Land Use Policy. 2020;91:104–382. doi: 10.1016/j.landusepol.2019.104382. DOI

Feng M, Bai Y. A global land cover map produced through integrating multi-source datasets. Big Earth Data. 2019;3(3):191–219. doi: 10.1080/20964471.2019.1663627. DOI

Feranec J, Soukup T, Hazeu G, Jaffrain G. European landscape dynamics: CORINE land cover data. CRC Press; Boca Raton: 2016.

Foley JA, DeFries R, Asner GP, Barford C, Bonan G, Carpenter SR, Chapin FS, Coe MT, Daily GC, Gibbs HK. Global consequences of land use. Science. 2005;309(5734):570–574. doi: 10.1126/science.1111772. PubMed DOI

Gao B-C. NDWI—a normalized difference water index for remote sensing of vegetation liquid water from space. Remote Sensing of Environment. 1996;58(3):257–266. doi: 10.1016/S0034-4257(96)00067-3. DOI

Gao Y, Liu L, Zhang X, Chen X, Mi J, Xie S. Consistency analysis and accuracy assessment of three global 30-m land-cover products over the european union using the LUCAS dataset. Remote Sensing. 2020;12(21):3479. doi: 10.3390/rs12213479. DOI

Géron A. Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow: concepts, tools, and techniques to build intelligent systems. O’Reilly Media; 2019.

Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA, Tyukavina A, Thau D, Stehman S, Goetz SJ, Loveland TR. High-resolution global maps of 21st-century forest cover change. Science. 2013;342(6160):850–853. doi: 10.1126/science.1244693. PubMed DOI

Hengl T, Leal Parente L, Križan J, Bonannella C. Continental Europe digital terrain model at 30 m resolution based on GEDI, ICESat-2, AW3D, GLO-30, EUDEM, MERIT DEM and background layers. Zenodo. 2021 doi: 10.5281/zenodo.4724549. DOI

Herold M, Mayaux P, Woodcock C, Baccini A, Schmullius C. Some challenges in global land cover mapping: an assessment of agreement and accuracy in existing 1 km datasets. Remote Sensing of Environment. 2008;112(5):2538–2556. doi: 10.1016/j.rse.2007.11.013. DOI

Hillger D, Kopp T, Lee T, Lindsey D, Seaman C, Miller S, Solbrig J, Kidder S, Bachmeier S, Jasmin T. First-light imagery from Suomi NPP VIIRS. Bulletin of the American Meteorological Society. 2013;94(7):1019–1029. doi: 10.1175/BAMS-D-12-00097.1. DOI

Homer C, Dewitz J, Fry J, Coan M, Hossain N, Larson C, Herold N, McKerrow A, VanDriel JN, Wickham J. Completion of the 2001 national land cover database for the counterminous United States. Photogrammetric Engineering and Remote Sensing. 2007;73(4):337.

Homer C, Dewitz J, Jin S, Xian G, Costello C, Danielson P, Gass L, Funk M, Wickham J, Stehman S. Conterminous United States land cover change patterns 2001–2016 from the 2016 national land cover database. ISPRS Journal of Photogrammetry and Remote Sensing. 2020;162:184–199. doi: 10.1016/j.isprsjprs.2020.02.019. PubMed DOI PMC

Hong C, Burney JA, Pongratz J, Nabel JE, Mueller ND, Jackson RB, Davis SJ. Global and regional drivers of land-use emissions in 1961–2017. Nature. 2021;589(7843):554–561. doi: 10.1038/s41586-020-03138-y. PubMed DOI

Houghton RA, House JI, Pongratz J, Van Der Werf GR, DeFries RS, Hansen MC, Quéré CL, Ramankutty N. Carbon emissions from land use and land-cover change. Biogeosciences. 2012;9(12):5125–5142. doi: 10.5194/bg-9-5125-2012. DOI

Huete AR. A soil-adjusted vegetation index (SAVI) Remote Sensing of Environment. 1988;25(3):295–309. doi: 10.1016/0034-4257(88)90106-X. DOI

Inglada J, Vincent A, Arias M, Tardy B, Morin D, Rodes I. Operational high resolution land cover map production at the country scale using satellite image time series. Remote Sensing. 2017;9(1):95. doi: 10.3390/rs9010095. DOI

Jin S, Sader SA. Comparison of time series tasseled cap wetness and the normalized difference moisture index in detecting forest disturbances. Remote Sensing of Environment. 2005;94(3):364–372. doi: 10.1016/j.rse.2004.10.012. DOI

Kaplan JO, Krumhardt KM, Ellis EC, Ruddiman WF, Lemmen C, Goldewijk KK. Holocene carbon emissions as a result of anthropogenic land cover change. The Holocene. 2011;21(5):775–791. doi: 10.1177/0959683610386983. DOI

Key CH, Benson NC. Measuring and remote sensing of burn severity. Proceedings joint fire science conference and workshop, volume 2; Moscow, ID. 1999. p. 284.

Key CH, Benson NC. Landscape assessment (LA). FIREMON: fire effects monitoring and inventory system. Technical Report RMRS-GTR-164-CD, volume 164; Fort Collins, CO. 2006.

Kilibarda M, Hengl T, Heuvelink GB, Gräler B, Pebesma E, Perčec Tadić M, Bajat B. Spatio-temporal interpolation of daily temperatures for global land areas at 1 km resolution. Journal of Geophysical Research: Atmospheres. 2014;119(5):2294–2313. doi: 10.1002/2013JD020803. DOI

Liu H, Gong P, Wang J, Clinton N, Bai Y, Liang S. Annual dynamics of global land cover and its long-term changes from 1982 to 2015. Earth System Science Data. 2020a;12(2):1217–1243. doi: 10.5194/essd-12-1217-2020. DOI

Liu L, Zhang X, Gao Y, Chen X, Shuai X, Mi J. Finer-resolution mapping of global land cover: recent developments, consistency analysis, and prospects. Journal of Remote Sensing. 2021;2021:5289697. doi: 10.34133/2021/5289697. DOI

Liu X, Trogisch S, He J-S, Niklaus PA, Bruelheide H, Tang Z, Erfmeier A, Scherer-Lorenzen M, Pietsch KA, Yang B. Tree species richness increases ecosystem carbon storage in subtropical forests. Proceedings of the Royal Society B: Biological Sciences. 2018;285(1885):20181240. doi: 10.1098/rspb.2018.1240. PubMed DOI PMC

Liu Y, Hou X, Li X, Song B, Wang C. Assessing and predicting changes in ecosystem service values based on land use/cover change in the Bohai Rim coastal zone. Ecological Indicators. 2020b;111:106004. doi: 10.1016/j.ecolind.2019.106004. DOI

Lovelace R, Nowosad J, Muenchow J. Geocomputation with R. CRC Press; Boca Raton: 2019. (Chapman & Hall/CRC The R Series).

Malinowski R, Lewiński S, Rybicki M, Gromny E, Jenerowicz M, Krupiński M, Nowakowski A, Wojtkowski C, Krupiński M, Krätzschmar E. Automated production of a land cover/use map of Europe based on Sentinel-2 imagery. Remote Sensing. 2020;12(21):3523. doi: 10.3390/rs12213523. DOI

McCulloch WS, Pitts W. A logical calculus of the ideas immanent in nervous activity. The Bulletin of Mathematical Biophysics. 1943;5(4):115–133. doi: 10.1007/BF02478259. PubMed DOI

Palahi M, Valbuena R, Senf C, Acil N, Pugh TA, Sadler J, Seidl R, Potapov P, Gardiner B, Hetemäki L. Concerns about reported harvests in European forests. Nature. 2021;592(7856):E15–E17. doi: 10.1038/s41586-021-03292-x. PubMed DOI

Paulsson J, Claesson S, Fridman J, Olsson H. Incorrect figures on harvested forests in Nature article. https://www.slu.se/en/ew-news/2020/7/incorrect-figures-on-harvested-forests-in-nature-article/ [4 July 2021];SLU News. 2020

Payn T, Carnus J-M, Freer-Smith P, Kimberley M, Kollert W, Liu S, Orazio C, Rodriguez L, Silva LN, Wingfield MJ. Changes in planted forests and future global implications. Forest Ecology and Management. 2015;352:57–67. doi: 10.1016/j.foreco.2015.06.021. DOI

Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V. Scikit-learn: machine learning in Python. Journal of Machine Learning Research. 2011;12:2825–2830.

Pflugmacher D, Rabe A, Peters M, Hostert P. Mapping pan-European land cover using Landsat spectral-temporal metrics and the European LUCAS survey. Remote Sensing of Environment. 2019;221:583–595. doi: 10.1016/j.rse.2018.12.001. DOI

Picard N, Leban J-M, Guehl J-M, Dreyer E, Bouriaud O, Bontemps J-D, Landmann G, Colin A, Peyron J-L, Marty P. Recent increase in European forest harvests as based on area estimates (Ceccherini et al. 2020a) not confirmed in the French case. Annals of Forest Science. 2021;78(1):1–5. doi: 10.1007/s13595-020-01014-3. DOI

Pielke Sr RA, Marland G, Betts RA, Chase TN, Eastman JL, Niles JO, Niyogi DDS, Running SW. The influence of land-use change and landscape dynamics on the climate system: relevance to climate-change policy beyond the radiative effect of greenhouse gases. Philosophical Transactions of the Royal Society of London. Series a: Mathematical, Physical and Engineering Sciences. 2002;360(1797):1705–1719. doi: 10.1098/rsta.2002.1027. PubMed DOI

Potapov P, Hansen MC, Kommareddy I, Kommareddy A, Turubanova S, Pickens A, Adusei B, Tyukavina A, Ying Q. Landsat analysis ready data for global land cover and land cover change mapping. Remote Sensing. 2020;12(3):426. doi: 10.3390/rs12030426. DOI

Qi J, Chehbouni A, Huete AR, Kerr YH, Sorooshian S. A modified soil adjusted vegetation index. Remote Sensing of Environment. 1994;48(2):119–126. doi: 10.1016/0034-4257(94)90134-1. DOI

Raši R. Ministerial conference on the protection of forests in Europe. 2020. [5 October 2021]. State of Europes Forests 2020.

Riley SJ, DeGloria SD, Elliot R. Index that quantifies topographic heterogeneity. Intermountain Journal of Sciences. 1999;5(1-4):23–27.

Roberts DR, Bahn V, Ciuti S, Boyce MS, Elith J, Guillera-Arroita G, Hauenstein S, Lahoz-Monfort JJ, Schröder B, Thuiller W. Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. Ecography. 2017;40(8):913–929. doi: 10.1111/ecog.02881. DOI

Román MO, Wang Z, Sun Q, Kalb V, Miller SD, Molthan A, Schultz L, Bell J, Stokes EC, Pandey B. NASA’s Black Marble nighttime lights product suite. Remote Sensing of Environment. 2018;210:113–143. doi: 10.1016/j.rse.2018.03.017. DOI

Rossi F, Breidenbach J, Puliti S, Astrup R, Talbot B. Assessing harvested sites in a forested boreal mountain catchment through global forest watch. Remote Sensing. 2019;11(5):543. doi: 10.3390/rs11050543. DOI

Sala OE, Chapin FS, Armesto JJ, Berlow E, Bloomfield J, Dirzo R, Huber-Sanwald E, Huenneke LF, Jackson RB, Kinzig A. Global biodiversity scenarios for the year 2100. Science. 2000;287(5459):1770–1774. doi: 10.1126/science.287.5459.1770. PubMed DOI

Seabold S, Perktold J. statsmodels: econometric and statistical modeling with Python. 9th Python in science conference.2010.

Senf C, Pflugmacher D, Zhiqiang Y, Sebald J, Knorn J, Neumann M, Hostert P, Seidl R. Canopy mortality has doubled in Europes temperate forests over the last three decades. Nature Communications. 2018;9(1):1–8. doi: 10.1038/s41467-017-02088-w. PubMed DOI PMC

Senf C, Seidl R. Mapping the forest disturbance regimes of Europe. Nature Sustainability. 2021;4(1):63–70. doi: 10.1038/s41893-020-00609-y. DOI

Seni G, Elder J. Ensemble methods in data mining: improving accuracy through combining predictions. Morgan & Claypool Publishers; San Rafael: 2010. (Synthesis lectures on data mining and knowledge discovery).

Shahi K, Shafri HZ, Taherzadeh E, Mansor S, Muniandy R. A novel spectral index to automatically extract road networks from WorldView-2 satellite imagery. The Egyptian Journal of Remote Sensing and Space Science. 2015;18(1):27–33. doi: 10.1016/j.ejrs.2014.12.003. DOI

Shumba T, De Vos A, Biggs R, Esler KJ, Ament JM, Clements HS. Effectiveness of private land conservation areas in maintaining natural land cover and biodiversity intactness. Global Ecology and Conservation. 2020;22:e00935. doi: 10.1016/j.gecco.2020.e00935. DOI

Song X-P, Hansen MC, Stehman SV, Potapov PV, Tyukavina A, Vermote EF, Townshend JR. Global land change from 1982 to 2016. Nature. 2018;560(7720):639–643. doi: 10.1038/s41586-018-0411-9. PubMed DOI PMC

Sy S, Quesada B. Anthropogenic land cover change impact on climate extremes during the 21st century. Environmental Research Letters. 2020;15(3):034002. doi: 10.1088/1748-9326/ab702c. DOI

Szantoi Z, Geller GN, Tsendbazar N-E, See L, Griffiths P, Fritz S, Gong P, Herold M, Mora B, Obregón A. Addressing the need for improved land cover map products for policy support. Environmental Science & Policy. 2020;112:28–35. doi: 10.1016/j.envsci.2020.04.005. PubMed DOI PMC

Townshend JR, Masek JG, Huang C, Vermote EF, Gao F, Channan S, Sexton JO, Feng M, Narasimhan R, Kim D. Global characterization and monitoring of forest cover using Landsat data: opportunities and challenges. International Journal of Digital Earth. 2012;5(5):373–397. doi: 10.1080/17538947.2012.713190. DOI

Trenberth KE. What are the seasons? Bulletin of the American Meteorological Society. 1983;64(11):1276–1282. doi: 10.1175/1520-0477(1983)064<1276:WATS>2.0.CO;2. DOI

Trisurat Y, Shirakawa H, Johnston JM. Land-use/land-cover change from socio-economic drivers and their impact on biodiversity in Nan Province, Thailand. Sustainability. 2019;11(3):649. doi: 10.3390/su11030649. PubMed DOI PMC

Tsendbazar N, Herold M, De Bruin S, Lesiv M, Fritz S, Van De Kerchove R, Buchhorn M, Duerauer M, Szantoi Z, Pekel J-F. Developing and applying a multi-purpose land cover validation dataset for Africa. Remote Sensing of Environment. 2018;219:298–309. doi: 10.1016/j.rse.2018.10.025. DOI

Tucker CJ. Red and photographic infrared linear combinations for monitoring vegetation. Remote Sensing of Environment. 1979;8(2):127–150. doi: 10.1016/0034-4257(79)90013-0. DOI

Van Rijsbergen C. Information retrieval. Butterworth Heinemann; Oxford: 1980.

Van Thinh T, Cao Duong P, Nishida Nasahara K. How does land use/land cover map’s accuracy depend on number of classification classes? SOLA. 2019;15:28–31. doi: 10.2151/sola.2019-006. DOI

Veldkamp A, Lambin EF. Predicting land-use change. Agriculture Ecosystems and Environment. 2001;85:1–6.

Venter ZS, Sydenham MA. Continental-scale land cover mapping at 10 m resolution over Europe (ELC10) Remote Sensing. 2021;13(12):2301. doi: 10.3390/rs13122301. DOI

Vilar L, Garrido J, Echavarra P, Martnez-Vega J, Martn M. Comparative analysis of CORINE and climate change initiative land cover maps in Europe: implications for wildfire occurrence estimation at regional and local scales. International Journal of Applied Earth Observation and Geoinformation. 2019;78:102–117. doi: 10.1016/j.jag.2019.01.019. DOI

Wang X, Wu C, Peng D, Gonsamo A, Liu Z. Snow cover phenology affects alpine vegetation growth dynamics on the Tibetan Plateau: satellite observed evidence, impacts of different biomes, and climate drivers. Agricultural and Forest Meteorology. 2018;256:61–74.

Zhang C, Ma Y. Ensemble machine learning: methods and applications. Springer New York; New York: 2012.

Zhang X, Liu L, Chen X, Gao Y, Xie S, Mi J. GLC_FCS30: global land-cover product with fine classification system at 30 m using time-series Landsat imagery. Earth System Science Data Discussions. 2020;13:2753–2776. doi: 10.5194/essd-13-2753-2021. DOI

Zhou W, Qian Y, Li X, Li W, Han L. Relationships between land cover and the surface urban heat island: seasonal variability and effects of spatial and thematic resolution of land cover data on predicting land surface temperatures. Landscape Ecology. 2014;29(1):153–167. doi: 10.1007/s10980-013-9950-5. DOI