Czech User Forum 3

The aim of precision agriculture is an optimization of production inputs (fertilizers, pesticides, fuel, etc.) based on the site specific crop requirements and soil condition. Crop management in this way can lead to economically efficient use of agrochemicals and minimization of environmental risks. 

Application of precision agriculture technologies in crop production requires high quality input data, which allows to delineate management zones within the fields. The preparation of these data is based on detailed mapping of soil heterogeneity and assessment of the current state crop stands by using traditional techniques, remote sensing and proximal mapping with various sensors. However, processing and management of all mentioned data is for farmers not simple and usually discourages users from the practice of comprehensive evaluation. Within the European project “Farm-Oriented Open Data in Europe” (FOODIE, 7FP No. 621074) is being developed cloud-based platform for collecting, managing and analyzing spatial data, allowing the integration of open data (LPIS, remote sensing and other) with farm records and data from sensor measurements (machinery tracking, agrometeorological monitoring). It can be assumed that the integration of all available information from the farm and their processing using GIS tools could identify the critical areas of land management, which are used for decision support of farmers and agronomists. 

The Forest Management Institute is currently focusing on the development of services providing information on forest resources based on remote sensing data with special interest in Sentinel-2 satellite optical data. Our methodology for forest health assessment is based on time series analysis of vegetation indices derived from (raw) Sentinel-2 satellite data. Vegetation index is a simple mathematical combination of two or more satellite image bands, which are sensitive to particular feature of interest (e.g. forest health status reflected via volume of growing stock, amount of defoliation, chlorophyll content). Sentinel-2 data will offer observation in several spectral channels sensitive to both biomass (e.g. assessing the contrast between red and near-infrared spectral region) and the photochemical pigments (“red edge” spectral region on the transition between red and near-infrared wavelengths). Thanks to the high revisit time, Sentinel-2 data will allow the normalization of the observations for equal time period (typically the peak growing season). Due to frequent observations we will be able to work with high-quality (cloud free) optical satellite observations. Analysis of the time series of vegetation indices can be used to study both long term trends in vegetation health status (cumulative sum of the vegetation index, slope of linear regression) and rapid health status decrease (anomaly detection in vegetation index time series). When studying the relative differences in the vegetation indices in time instead of their instantaneous absolute values, we will be able to overcome the species and structure-specificity of vegetation indices. The outputs of the service (e.g. the mapserver) should serve both to streamline the implementation of forestry-related measures and for redistribution of subsidies and tax relief for forest owners, and ultimately for the forest monitoring activities.

The risk of fires in the countryside is growing due to climate change and land cover change in the Czech Republic. Fire danger depends on the weather and the intensity of drought. This paper illustrates the potential utilization of anomalies in Fraction of Absorbed Photosynthetically Active Radiation (fAPAR) for assessment of drought. The European Space Agency (ESA) is preparing fAPAR from data collected by the MERIS instrument onboard the Envisat satellite. The fAPAR values were used to determine the onset beginning of phenological phases, tracking of seasonal changes of green vegetation growing and timing of key phenological events. All of the dry periods that were detected by the Standardized Precipitation Evapotranspiration Index (SPEI) and recorded number of wildfires correlated well with anomalies of fAPAR. A decrease in fAPAR correlation with SPEI was only observed at high elevations. Our results demonstrate that fAPAR values may be used for assessment of drought within the vegetation fire danger forecasts in the Central European region.

Soil erosion is a one of the major degradation factors of soil worldwide. In the Czech Republic, area about 50 % of agricultural land is endangered by water erosion and about 10 % by wind erosion. 

Nevertheless, up to now there have not been sufficient information on widespread effect of soil erosion. In the Czech republic, there are specific conditions, which contribute to the soil degradation issue. These conditions include hilly landscape, large units of arable land, lack of landscape lines and barriers that prevent surface runoff or using inappropriate agriculture methods.

The soil erosion issue is currently solved by detailed survey on particular fields or by numerical and empirical erosion models on a regional scale. However, the models allow only for predicting the potential of soil erosion. A large scale assessment of the current state of soil degradation caused by erosion is thus based solely on expert knowledge. In this regard, remote sensing methods offer the possibility to monitor the real state of soil degradation caused by erosion on different scales.

The use of available remote sensing data acquired at short periodic intervals is most suitable for this purpose, particularly the data from satellites Landsat 8 and Sentinel 2. Determining the degree of soil degradation caused by erosion is possible by soil properties distribution assessment acquired through image spectral data analysis. The analysis should be carry out preferably on the locations in their optimal conditions. The conditions constitute dry, bare (without crops or crop residues), flat (prepared for sowing) soil scanned on cloudless image. Virtual constellation of both satellites allows for capturing these required conditions within a few consecutive years.

Aerial hyperspectral images (data with very good spatial and spectral resolution) have been used for the analysis of applicability of remote sensing data for assessment of extent of soil degradation caused by erosion. The data from the CASI, SASI and the TASI sensors has been acquired and preprocessed using laboratory FLIS (Flying Laboratory of Imaging Systems), Global Change Research Institute CAS. The data has been further processed using the acquired field soil and morphometric data applying advanced statistical regression methods (PLSR, SVMR). The result is predictive maps of surface soil properties and their redistribution, from which the information on soil degradation can be obtained. The predictive ability of this method varies depending on the statistical method, and especially on the degree of heterogeneity of the soil properties and the variability of the soil parent material in the surveyed area. The analysis has proved that the same method can be used using satellite data, albeit with less spatial resolution and a slightly worse prediction ability of statistical models.

The Remote sensing department at the Czech Geological Survey (SGS) has a long-term experience with employing remote sensing methods to mineral mapping, soil modeling and assessment of mining impacts. The presentation will introduce the Czech remote sensing community with international activities which has been carried at the CGS. The focus will be on demonstration of Copernicus data utilization as well as on presenting international activities within the EC projects. Special emphasis will be put on mineral mapping when using simulated Sentinel 2 and EnMap data provided by the Helmholtz Centre Potsdam (GFZ: German Research Centre for Geosciences).