Machine learning in modelling land-use and land cover-change (LULCC): Current status, challenges and prospects.

Land-use and land-cover change (LULCC) are of importance in natural resource management, environmental modelling and assessment, and agricultural production management. However, LULCC detection and modelling is a complex, data-driven process in the remote sensing field due to the processing of massive historical and current data, real-time interaction of scenario data, and spatial environmental data. In this paper, we review principles and methods of LULCC modelling, using machine learning and beyond, such as traditional cellular automata (CA). Then, we examine the characteristics, capabilities, limitations, and perspectives of machine learning. Machine learning has not yet been dramatic in modelling LULCC, such as urbanization prediction and crop yield prediction because competition and transition between land cover types are dynamic at a local scale under varying natural drivers and human activities. Upcoming challenges of machine learning in modelling LULCC remain in the detection and prediction of LULC evolutionary processes if considering their applicability and feasibility, such as the spatio-temporal transition mechanisms to describe occurrence, transition, spreading, and spatial patterns of changes, availability of training data of all the change drivers, particularly sequence data, and identification and inclusion of local ecological, hydrological, and social-economic drivers in addressing the spectral feature change. This review points out the need for multidisciplinary research beyond image processing and pattern recognition of machine learning in accelerating and advancing studies of LULCC modelling. Despite this, we believe that machine learning has strong potentials to incorporate new exploratory variables in modelling LULCC through expanding remote sensing big data and advancing transient algorithms.

Wang J ，Bretz M ，Dewan MAA ，Delavar MA ... -  《-》

Integrating stakeholders' perspectives and spatial modelling to develop scenarios of future land use and land cover change in northern Tanzania.

Kariuki RW ，Munishi LK ，Courtney-Mustaphi CJ ，Capitani C ，Shoemaker A ，Lane PJ ，Marchant R ... -  《PLoS One》

Land use land cover change and the resilience of social-ecological systems in a sub-region in South west Cameroon.

Ewane EB 《-》

Assessing driving forces of land use and land cover change by a mixed-method approach in north-eastern Ghana, West Africa.

Land use and land cover change (LULCC) is the result of complex human-environmental interactions. The high interdependencies in social-ecological systems make it difficult to identify the main drivers. However, knowledge of key drivers of LULCC, including indirect (underlying) drivers which cannot be easily determined by spatial or economic analyses, is essential for land use planning and especially important in developing countries. We used a mixed-method approach in order to detect drivers of LULCC in the Upper East Region of northern Ghana by different qualitative and quantitative methods which were compared in a confidence level analysis. Viewpoints from experts help to answer why the land use is changing, since many triggering effects, especially non-spatial and indirect drivers of LULCC, are not measurable by other methodological approaches. Geo-statistical or economic analyses add to validate the relevance of the expert-based results. First, we conducted in-depth interviews and developed a list of 34 direct and indirect drivers of LULCC. Subsequently, a group of experts was asked in a questionnaire to select the most important drivers by using a Likert scale. This information was complemented by remote sensing analysis. Finally, the driver analysis was compared to information from literature. Based on these analyses there is a very high confidence that population growth, especially in rural areas, is a major driver of LULCC. Further, current farming practice, bush fires, livestock, the road network and climate variability were the main direct drivers while the financial capital of farmers and customary norms regarding land tenure were listed as important indirect drivers with high confidence. Many of these driving forces, such as labour shortage and migration, are furthermore interdependent. Governmental laws, credits, the service by extension officers, conservational agriculture and foreign agricultural medium-scale investments are currently not driving land use changes. We conclude that the mixed-method approach improves the confidence of findings and the selection of most important drivers for modelling LULCC, especially in developing countries.

Kleemann J ，Baysal G ，Bulley HNN ，Fürst C ... -  《-》

Land use land cover change in the African Great Lakes Region: a spatial-temporal analysis and future predictions.

The African Great Lakes Region has experienced substantial land use land cover change (LULCC) over the last decades, driven by a complex interplay of various factors. However, a comprehensive analysis exploring the relationships between LULCC, and its explanatory variables remains unexplored. This study focused on the Lake Kivu catchment in Rwanda, analysing LULCC from 1990 to 2020, identifying major variables, and predicting future LULC scenarios under different development trajectories. Image classification was conducted in Google Earth Engine using random forest classifier, by incorporating seasonal composites Landsat images, spectral indices, and topographic features, to enhance discrimination and capture seasonal variations. The results demonstrated an overall accuracy exceeding 83%. Historical analysis revealed significant changes, including forest loss (26.6 to 18.7%) and agricultural land expansion (27.7 to 43%) in the 1990-2000 decade, attributed to political conflicts and population movements. Forest recovery (24.8% by 2020) was observed in subsequent decades, driven by Rwanda's sustainable development initiatives. A Multi-Layer Perceptron neural network from Land Change Modeler predicted distinct 2030 and 2050 LULC scenarios based on natural, socio-economic variables, and historical transitions. Analysis of explanatory variables highlighted the significant role of proximity to urban centers, population density, and terrain in LULCC. Predictions indicate distinct trajectories influenced by demographic and socio-economic trends. The study recommends adopting the Green Growth Economy scenario aligned with ongoing conservation measures. The findings contribute to identifying opportunities for land restoration and conservation efforts, promoting the preservation of Lake Kivu catchment's ecological integrity, in alignment with national and global goals.

Kayitesi NM ，Guzha AC ，Tonini M ，Mariethoz G ... -  《-》