Spatially Stratified and Multi-Stage Approach for National Land Cover Mapping Based on Sentinel-2 Data and Expert Knowledge

Abstract

Portugal is building a land cover monitoring system to deliver land cover products annually for its mainland territory. This paper presents the methodology developed to produce a prototype relative to 2018 as the first land cover map of the future annual map series (COSsim). A total of thirteen land cover classes are represented, including the most important tree species in Portugal. The mapping approach developed includes two levels of spatial stratification based on landscape dynamics. Strata are analysed independently at the higher level, while nested sublevels can share data and procedures. Multiple stages of analysis are implemented in which subsequent stages improve the outputs of precedent stages. The goal is to adjust mapping to the local landscape and tackle specific problems or divide complex mapping tasks in several parts. Supervised classification of Sentinel-2 time series and post-classification analysis with expert knowledge were performed throughout four stages. The overall accuracy of the map is estimated at 81.3% (±2.1) at the 95% confidence level. Higher thematic accuracy was achieved in southern Portugal, and expert knowledge significantly improved the quality of the map.

Vegetation Mapping with Random Forest using Sentinel 2- A case study for Lousã region, Portugal

“…Sentinel-2 includes three red-edged vegetation and the SWIR bands that are highly susceptible to chlorophyll content and amend distinguishing different vegetation types and LCLU classification accuracy (Chaves et al 2020). Several studies found Sentinel-2 data with high potential in different applications such as crop classification (Hernandez et al 2020), tree species classification (Costa et al 2022, Wessel et al 2018, Persson et al 2018, mapping burned area (Pacheco et al 2021), and forest type classification (Kaplan 2021). Most of the recent studies have shown that non-parametric machine learning approaches such as Artificial Neural Network (ANN), Support Vector Machine (SVM), and Random Forest (RF) have a great potential to classify heterogeneous land covers (Wessel et al 2018, Pacheco et al 2021, Sheykhmousa et al 2020.…”

“…Usually, they cannot comply with the user-specific requirements in space and time (Costa et al 2018). Several researchers found the Mediterranean countries, such as Portugal and Spain, with the lowest overall accuracy, below 70% out of all European countries mapped (Costa et al 2022, Liu et al 2021. National land cover mapping with higher accuracy and up-to-date and detailed data can use as a complementary product for larger mapping scales (Costa et al 2022).…”

“…Several researchers found the Mediterranean countries, such as Portugal and Spain, with the lowest overall accuracy, below 70% out of all European countries mapped (Costa et al 2022, Liu et al 2021. National land cover mapping with higher accuracy and up-to-date and detailed data can use as a complementary product for larger mapping scales (Costa et al 2022). SMOS (Sistema de Monitorização de Ocupação do Solo) is the land cover monitoring system for mainland Portugal which is in preparation by Directorate General for Territory (DGT) based on COS (Cartografia de Uso e Ocupação do Solo), the traditional LCLU map of Portugal.…”

“…SMOS (Sistema de Monitorização de Ocupação do Solo) is the land cover monitoring system for mainland Portugal which is in preparation by Directorate General for Territory (DGT) based on COS (Cartografia de Uso e Ocupação do Solo), the traditional LCLU map of Portugal. An annual land cover cartography product of mainland Portugal (COSsim) based on sentinel-2 was started to overcome the limitation of COS. Costa et al (Costa et al 2022) presented an approach to map COSsim for 2018 with an overall accuracy of 81.3% using Random Forest (RF) classification and Sentinel-2 multi-temporal data. Although they achieved an accuracy of over 90% for southern Portugal, the classification overall accuracy for the center and North Portugal was below 80%, especially for the Coimbra district with approximately 75%.…”

“…Their methodology was developed based on the following stages such as image classification, spatial stratification, knowledge-based rules (combination of expert knowledge and auxiliary data), intra-annual change, and manual editing. They classified this map into 13 thematic classes including the predominant species such as evergreen oaks, Eucalyptus, Maritime pine, and Stone pine (Costa et al 2022). Costa et al (2018) also studied a methodology with the combination of supervised and rule-based classification under the scope of the Land Use and Coverage Area frame Survey (LUCAS) to produce annual land cover statistics over Portugal from 2010 to 2015.…”