TY - JOUR

T1 - A multi-sensor and multi-temporal remote sensing approach to detect land cover change dynamics in heterogeneous urban landscapes

AU - Kabisch, Nadja

AU - Selsam, Peter

AU - Kirsten, Toralf

AU - Lausch, Angela

AU - Bumberger, Jan

N1 - Funding information: The activities were co-financed by the research project Environmental-Health Interactions in Cities (GreenEquityHEALTH) – Challenges for Human Well-Being under Global Changes (2017–2022), funded by the German Federal Ministry of Education and Research (BMBF; no. 01LN1705A ) and the research project Smart Sensor-based Digital Ecosystem Services (S2DES, 2016-2020), funded by the European Social Fund (ESF; Grant Agreement No. 100269858 ). For supporting the development of automated data technologies these activities have received funding under H2020-SC5-15-2015 “Strengthening the European Research Area in the domain of Earth Observation” within the project “GEOEssentials” (ERA-NET-Cofund Grant, Grant Agreement No. 689443). The data from the Planet Labs company (order for RapidEye images, ref. Planet*, Planet.org*) were obtained from the UFZ contract (Tereno Contract no. 462/703).

PY - 2019/4

Y1 - 2019/4

N2 - With global changes such as climate change and urbanization, land cover is prone to changing rapidly in cities around the globe. Urban management and planning is challenged with development pressure to house increasing numbers of people. Most up-to date continuous land use and land cover change data are needed to make informed decisions on where to develop new residential areas while ensuring sufficient open and green spaces for a sustainable urban development. Optical remote sensing data provide important information to detect changes in heterogeneous urban landscapes over long time periods in contrast to conventional approaches such as cadastral and construction data. However, data from individual sensors may fail to provide useful images in the required temporal density, which is particularly the case in mid-latitudes due to relatively abundant cloud coverage. Furthermore, the data of a single sensor may be unavailable for an extended period of time or to the public at no cost. In this paper, we present an integrated, standardized approach that aims at combining remote sensing data in a high resolution that are provided by different sensors, are publicly available for a long-term period of more than ten years (2005–2017) and provide a high temporal resolution if combined. This multi-sensor and multi-temporal approach detects urban land cover changes within the highly dynamic city of Leipzig, Germany as a case. Landsat, Sentinel and RapidEye data are combined in a robust and normalized procedure to offset the variation and disturbances of different sensor characteristics. To apply the approach for detecting land cover changes, the Normalized Difference Vegetation Index (NDVI) is calculated and transferred into a classified NDVI (Classified Vegetation Cover – CVC). Small scale vegetation development in heterogeneous complex areas of a European compact city are highlighted. Results of this procedure show successfully that the presented approach is applicable with divers sensors’ combinations for a longer time period and thus, provides an option for urban planning to update their land use and land cover information timely and on a small scale when using publicly available no cost data.

AB - With global changes such as climate change and urbanization, land cover is prone to changing rapidly in cities around the globe. Urban management and planning is challenged with development pressure to house increasing numbers of people. Most up-to date continuous land use and land cover change data are needed to make informed decisions on where to develop new residential areas while ensuring sufficient open and green spaces for a sustainable urban development. Optical remote sensing data provide important information to detect changes in heterogeneous urban landscapes over long time periods in contrast to conventional approaches such as cadastral and construction data. However, data from individual sensors may fail to provide useful images in the required temporal density, which is particularly the case in mid-latitudes due to relatively abundant cloud coverage. Furthermore, the data of a single sensor may be unavailable for an extended period of time or to the public at no cost. In this paper, we present an integrated, standardized approach that aims at combining remote sensing data in a high resolution that are provided by different sensors, are publicly available for a long-term period of more than ten years (2005–2017) and provide a high temporal resolution if combined. This multi-sensor and multi-temporal approach detects urban land cover changes within the highly dynamic city of Leipzig, Germany as a case. Landsat, Sentinel and RapidEye data are combined in a robust and normalized procedure to offset the variation and disturbances of different sensor characteristics. To apply the approach for detecting land cover changes, the Normalized Difference Vegetation Index (NDVI) is calculated and transferred into a classified NDVI (Classified Vegetation Cover – CVC). Small scale vegetation development in heterogeneous complex areas of a European compact city are highlighted. Results of this procedure show successfully that the presented approach is applicable with divers sensors’ combinations for a longer time period and thus, provides an option for urban planning to update their land use and land cover information timely and on a small scale when using publicly available no cost data.

KW - Classified Vegetation Cover (CVC)

KW - Greenness

KW - Leipzig

KW - Multi-sensor

KW - Multi-temporal

KW - NDVI

KW - New approach

KW - Remote sensing

KW - Urban areas

UR - http://www.scopus.com/inward/record.url?scp=85058950501&partnerID=8YFLogxK

U2 - 10.1016/j.ecolind.2018.12.033

DO - 10.1016/j.ecolind.2018.12.033

M3 - Article

AN - SCOPUS:85058950501

VL - 99

SP - 273

EP - 282

JO - Ecological indicators

JF - Ecological indicators

SN - 1470-160X

ER -