An Algorithm for Maximum Concordance to Harmonize Legends of Land Use and Land Cover Maps
Land use and land cover maps (LULC) are abstractions of the physical space of a chosen region. Comparison of LULC maps is essential to understand landscape dynamics, alteration patterns, and environmental implications. This article has the objective of propose an algorithm for harmonizing LULC maps...
| Authors: | , , , |
|---|---|
| Format: | article |
| Status: | Published version |
| Publication Date: | 2025 |
| Country: | Brasil |
| Institution: | Universidade Federal de Uberlândia (UFU) |
| Repository: | Revista brasileira de cartografia - RBC (Online) |
| Language: | English |
| OAI Identifier: | oai:ojs.www.seer.ufu.br:article/72784 |
| Online Access: | https://seer.ufu.br/index.php/revistabrasileiracartografia/article/view/72784 |
| Access Level: | Open access |
| Keyword: | Harmonization algorithm National Inventory MapBiomas Land use and land cover maps harmonization algorithm national inventory land land use and land cover maps |
| Summary: | Land use and land cover maps (LULC) are abstractions of the physical space of a chosen region. Comparison of LULC maps is essential to understand landscape dynamics, alteration patterns, and environmental implications. This article has the objective of propose an algorithm for harmonizing LULC maps based on the spatial distribution of their classes and applies it in a case study to harmonize the maps of Brazil’s National Inventory of Anthropogenic Emissions by Sources and Removals of Greenhouse Gases (Fourth Version) and MapBiomas (Collection 7) based on their spatial distribution of LULC classes. The purpose of this paper is to compute the agreement between two initiatives. Furthermore, the results highlight the classes and areas of potential inconsistency or ambiguity, allowing to identify and correct discrepancies, proposing a harmonized legend between then. At the national level, we reached maximum agreement 81% between the two maps. Of the 44 equivalences, the algorithm accurately recognized 36 of the connections between the classes. At the biome level, the algorithm achieved its highest concordance within the Amazonia biome, surpassing Brazil’s level by 11%, mainly due to the size and homogeneity of the forest classes. In biomes with a predominance of nonforest vegetation, an increased confusion was observed among the classes ‘Grassland’, ‘Pasture’, and ‘Forest’ was observed between the maps, especially in Pampa and Caatinga. |
|---|