The use of remote sensing methods has a long tradition at NLI which dates back to the 1980s. A significant development in this area was connected with the second cycle of the National Forest Inventory. In 2009, a specialised remote sensing workplace was established at the Frýdek-Místek branch and its professional and processing scope is continuously expanding.
We use multiple data sources to monitor the status and development of forests. These include aerial surveys (LMS) from the national aerial photography organised by the Czech Office for Surveying, Mapping and Cadastre (ČÚZK). Within the European COPERNICUS programme we process satellite imagery from Sentinel satellites (in particular SENTINEL-2 multispectral satellite data), long-term series are analysed from the LANDSAT programme, the latest source used is PlanetScope commercial satellite data with high spatial and temporal resolution.
The applicability of the different methods and data sources depends on the size of the monitored area and the requirement for data timeliness. The advantage of satellite data is the potential for repeated monitoring of large areas (e.g., the whole country) with a short repeat period: annually / several times per growing season / after several days / daily / up to several times per day. The advantage of LMS is its high positional accuracy (mean positional error < 1 m) and high spatial resolution (currently < 20 cm per pixel) with repeated imaging of half of the Czech Republic annually (i.e., full coverage every 2 years). For nationwide survey of forest condition, the optimal approach is a combination of remote sensing and ground measurement, on which, for example, the National Forest Inventory in the Czech Republic is based.
The primary use of LMS are analyses for further refinement of the NFI using an extended network of inventory points and transects surveyed by optical stereophotogrammetry and repeated identification of areas for field survey in a given campaign season. At the same time, a training dataset is established, which is further used for automated image classification using self-learning algorithms. A unique output of the LMS analysis is a normalized digital surface model (nDSM) that displays stand heights at 1 m spatial resolution at the date of LMS acquisition. Based on the nDSM, we derive, for example, the Growth Phase Map, the Stand Segment Map and the Harvest Detection. Another output is a mosaic in the infrared spectrum (CIR orthophoto), which in false colours gives a much better view of the vegetation status, e.g., species of tree species, presence of droughts, etc. We are also creating a Forest Edge Map according to NFI (OLIL) on top of the LMS images. All outputs are regularly updated according to the national aerial photography cycle.
Satellite data provides systematic observations of any place on the earth. The use of these data is of great benefit for monitoring the current status and changes in the landscape, as they allow the repeated mapping of large areas, even several times a year. Compared to LMS, satellite data have a lower spatial resolution of tens of metres and do not allow individual trees to be tracked. Their strength is their ability to image in several regions of infrared light, which is invisible to the human eye and shows some indicators of forest health. Also, very valuable for us are the archival satellite images, which in the case of LANDSAT date back to 1984. Thanks to them, we are able to assess, for example, long-term trends in forest health or changes in forest species composition. The combination of current Sentinel-2 satellite data capturing forest stands in different phenophases and NIL ground survey data allows us to produce a Tree Map for the main economic tree species of the Czech Republic, and the analysis of PlanetScope data at high spatial resolution is behind the successful Bark Map project.
In addition to internal use at NLI, our outputs serve other institutions, e.g., Ministry of Agriculture (MZe), The State Agricultural Intervention Fund (SZIF), the Czech Office for Surveying, Mapping and Cadastre (ČÚZK), Forests of Czech Republic (LČR), state forest administration and various academic and scientific institutions.