Efficient Earth Observation Handling and Computation.
Quantification of spatial-temporal variability of carbon fluxes, carbon stocks and vegetation phenology are key questions targeting effects of climatic change, resource availability and the connectivity between climatic drivers and environmental responses. Large research efforts involving natural science departments at Lund and elsewhere targets these questions which include handling and processing of large amounts of geospatial data, originating from Earth Observation (EO) satellites, airplanes and unmanned aerial vehicles (UAVs), but also climate data, climate scenarios, LIDAR data etc. Our studies concern extraction of relevant information from these data, including estimates of carbon fluxes in terrestrial ecosystems drought monitoring, studies of vegetation phenology in tropical, boreal and arctic environments , land use and land cover change, often quantified through time series analysis and trend analysis of EO data. Recently, the amount of freely available EO data has increased significantly, through the global LANDSAT archive, including >5 million satellite images equivalent to >4100 Tb of data, and through new EO satellites. In August 2015 was SENTINEL-2 launched, a new optical EO system hosted by ESA and an important part of COPERNICUS, EU’s environmental monitoring program. This EO system will cover the globe every 3-5 days with a spatial resolution of 10-60 meters and 16 spectral bands, providing a unique range of monitoring possibilities and 2 TB of data daily, complementary to existing data sources. Recent cooperation with NASA provide access to hundreds of TB high spatial resolution (50 cm) commercial data. Our studies of vegetation phenology, carbon fluxes and carbon stocks are of high international quality. Combining these scientific questions, the big data available, and our analytical tools with increased storing and computing capabilities will boost our scientific impact within the following research activities: a) Development of “state of the art” software for time series analysis of EO data is since 15 years an important part of our research, constituting the TIMESAT package, providing efficient handling and analysis of time series of EO data. Increased computational power needed. b) analysis of data collected from UAVs at high spatial resolution. Through the recently formed NordSpec project we co-ordinate UAV-based spectral data collection for the a network of Swedish ecological research stations across Sweden. Resulting image data has cm-resolution and contain very large amounts of data placing high demand on processing capability and data storage. Supported by FORMAS 2017-2019. c) Multi-resolution assessment of land degradation: This project is cooperative including Lund University and NASA (Prof. Tucker) and addresses methods to estimate status and trends in land degradation through use of multiresolution EO data. NASA will freely provide commercial satellite data for use in conjunction with NASA research projects through their recent NGA Commercial Archive Data initiative. Hence thousands of commercial satellite images from subsaharan Africa be available for the time period of 2007-2015. d) Emerging studies in epidemiology integrating temporal-spatial statistics derived from EO data with traditionally health related statistics. Supported by VR 2017-2020.