The evolution of desertification modelling research: A bibliometric analysis of thematic structure and change

Desertification is a long-standing and growing challenge in dryland regions worldwide, with serious consequences for ecosystems and livelihoods. Because desertification often results from an interaction between climatic variability, land use practices, and socio-economic pressures, models have become essential tools to analyse processes, explore future scenarios, and support environmental decision-making. Over the past four decades, a wide range of models has been applied to desertification research. However, it remains unclear how this body of modelling work has evolved over time, what it has prioritised, and whether it adequately supports land management and policy interventions.

To address this gap, a bibliometric analysis focusing on thematic structure and evolution was conducted to examine desertification modelling research published between 1981 and 2026. A broad search of the Web of Science database was used to capture studies linking desertification, and modelling, which after removing duplicates, resulted in a dataset of 3,200 scientific publications. Using the bibliometrix package in R, publication trends and keyword relationships were analysed to identify dominant research topics and their development over time. Thematic maps were produced for four consecutive periods to assess which themes are most central to the field and how conceptually developed they are within each period. A thematic evolution analysis was then applied to summarise how major research themes persist, emerge, or decline across decades.

The results show that desertification modelling research has expanded substantially across the study period, rising from a small number of studies in the 1980s and 1990s to rapid growth after 2000, with the majority of publications appearing in the last decade. Themes related to drought, rainfall variability, climate change, and modelling techniques consistently dominate the research landscape. On the other hand, impact and management themes remain marginal, suggesting a limited integration of social and decision-making dimensions within dominant modelling approaches. This observation was reinforced by the keyword co-occurrence network which showed an abundance of climate-related and modelling terms, while rarely showing social and management-oriented keywords. At the same time, desertification itself becomes a more specialised topic over time shifting towards a peripheral position relative to climate-driven and model-focused research.

These findings point to a clear imbalance when viewed through the Driver–Pressure–State–Impact–Response (DPSIR) framework, which conceptualises environmental problems as chains linking causes, system changes, consequences, and societal responses. While drivers and environmental states are well represented in the modelling literature, responses are weakly addressed. Building on this bibliometric analysis, the next phase of the study will undertake a scoping review of desertification models to systematically examine which DPSIR components are represented, how they are modelled, and to what extent existing models support management and policy decisions.