Small forest patch conservation elevates simulated configuration indicators in Kenyan drylands

Researchers have used scenario simulations to show that preserving small forest patches in arid and semi-arid landscapes (ASALs) in southern Kenya produces stronger spatial configuration indicators than large-scale forest consolidation or bundled management strategies. Published in Environmental and Sustainability Indicators, the study utilizes a calibrated spatially explicit land-use model to simulate how alternative land tenure regimes and regenerative management strategies reorganize landscape configuration in Narok County. The findings indicate that targeted interventions, specifically the retention of small forest patches, generate higher levels of spatial heterogeneity and functional adjacency, which are diagnostic signals for regenerative capacity in climate-vulnerable drylands.

The study addresses a significant gap in sustainability science, where existing indicator frameworks often prioritize land-cover extent or vegetation condition while underrepresenting landscape configuration. Configuration refers to the spatial arrangement, fragmentation, and adjacency of land uses, which regulate ecological flows and adaptive responses. In ASALs, fine-scale spatial relationships, such as proximity to remnant vegetation or water access, disproportionately influence buffering capacity and livelihood resilience under climate stress. The authors argue that translating these spatial reorganizations into interpretable, policy-relevant diagnostics requires configuration-based indicators that explicitly capture fragmentation, connectivity, and functional adjacency.

The research focuses on Narok County in southern Kenya, a region characterized by sharp ecological gradients from humid highlands to arid rangelands. This area supports intertwined pastoral, agropastoral, and farming livelihoods and reflects broader policy transitions in sub-Saharan African rangelands, including shifts from communal tenure to group ranches and subsequent privatization. These governance legacies interact with intensifying ecological pressures, such as erratic rainfall, to reshape landscape patterns. The study uses a previously calibrated Land Change Modeler (LCM) that achieved approximately 83% overall correctness against a 2018 reference map, providing a robust basis for scenario simulations.

The scenario design contrasts two land tenure regimes: group ranch governance and privatization. Under group ranch scenarios, historically collective areas remain under collective ownership, constraining cropland expansion within those zones. In contrast, privatization scenarios assume landscape-wide consolidation of individual tenure, promoting broader cropland expansion and higher rates of forest-to-cropland conversion. The researchers combined these tenure contexts with six regenerative management strategies, including small forest patch conservation, large forest patch conservation, forest connectivity enhancement, and strategic water-pan placement. This framework yielded 14 alternative scenario simulations in addition to a baseline prediction.

Results indicate that land tenure exerts a first-order influence on landscape configuration. In the absence of regenerative strategies, group ranch governance produced a more heterogeneous and better-connected mosaic than privatization. Relative to the baseline, the group ranch scenario increased forest cover by 2.3 percentage points and rangeland by 3.0 percentage points, while reducing cropland by 5.3 percentage points. Conversely, the privatization scenario reduced rangeland by 2.4 percentage points and slightly increased forest and cropland. Landscape-level edge density, an indicator of fine-scale heterogeneity, was substantially higher under group ranch tenure, reflecting a more spatially heterogeneous configuration.

Functional adjacency, which measures the spatial coupling between land uses, also varied significantly by tenure. Under group ranch governance, the proportion of cropland occurring within 500 meters of forest cover increased substantially compared to the baseline. This suggests a higher potential for edge-mediated functions such as pollination and microclimate buffering. In privatized landscapes, cropland-forest adjacency increased only slightly, reflecting more aggregated land use patterns. Rangeland-forest adjacency was also higher under group ranch tenure, indicating greater proximity of pastoral areas to forest refugia, which is critical for drought resilience and wildlife movement.

When regenerative management strategies were applied, their effectiveness was conditioned by the tenure context. Under group ranch governance, strategies prioritizing small forest patches or connectivity produced the largest increases in landscape-level edge density. Small-patch conservation consistently yielded the highest values for heterogeneity and adjacency indicators across both tenure regimes. In contrast, under privatization, most strategies produced edge-density values at or below the baseline, with only small-patch and connectivity strategies yielding slight increases. This attenuation suggests that parcelization and fencing under privatization constrain the spatial reorganization necessary for these strategies to be effective.

The study identified a “bundle paradox” where integrated multi-component strategy bundles produced weaker configuration responses than targeted interventions. Bundling small-patch conservation, large-patch protection, connectivity, and water-pan placement created offsetting configuration effects that reduced the expression of adjacency and heterogeneity indicators. This finding underscores the importance of specificity in landscape planning; targeted interventions that focus on patch dispersion and functional adjacency provide clearer diagnostic signals of regenerative potential than broad, aggregated packages.

Water-pan placement generated moderate but spatially concentrated indicator responses, particularly through increased rangeland-forest adjacency. By treating water pans as discrete resource nodes, the model showed how strategic siting could reorganize land-use proximity and support localized refugia during dry periods. This aligns with Kenya’s National Adaptation Plan, which prioritizes climate-resilient water infrastructure to stabilize pastoral livelihoods. However, the indicator gains were localized rather than system-wide, highlighting the need to balance water infrastructure with broader landscape connectivity efforts.

Sensitivity analysis revealed that the relative ranking of strategies remained stable across substantial perturbations of transition probabilities, ranging from ±25% to ±75%. Group ranch scenarios maintained higher values for edge density and adjacency indicators than privatization scenarios across all sensitivity levels. This robustness suggests that the observed differences reflect stable structural patterns rather than model-specific parameter choices. However, adjacency indicators showed greater sensitivity to perturbation than fragmentation metrics, indicating that localized spatial interactions are more variable and require careful comparative interpretation.

The authors emphasize that configuration-based indicators function as diagnostic signals of spatial structure rather than direct measures of ecological outcomes. While higher heterogeneity and connectivity create spatial conditions favorable for species movement and ecosystem services, the indicators do not quantify realized ecological benefits. Realized responses depend on factors such as vegetation condition, species-specific responses, and climate variability, which are not fully captured by land-cover configuration alone. Future work should pair these configuration indicators with field-based measures of ecosystem function to validate whether predicted spatial gains translate into tangible ecological and livelihood benefits.

This study contributes a transferable framework for indicator-based planning in multifunctional landscapes facing coupled climate and governance pressures. By treating spatial configuration as the intermediary between intervention and regenerative capacity, the approach offers a practical pathway for integrating spatial design and policy context. The findings suggest that sustainability assessments are best supported by focused indicators that reveal how specific design choices reorganize landscape structure, enabling adaptive, tenure-aware planning that can sustain multifunctionality and resilience in rapidly changing environments.

The study’s conceptual framework positions landscape configuration as the primary intermediary through which policy and management interventions influence regenerative landscape processes. This framework distinguishes between specific interventions, such as tenure policies and forest patch conservation, and the resulting configuration outcomes, which are measurable spatial pattern states. These states include heterogeneous mosaics characterized by high edge density, increased cross-habitat adjacency, and permeable networks produced by enhanced connectivity. The authors theorize that these configuration states mediate regenerative processes by shaping the flows of organisms, resources, and energy across the landscape. Such processes include edge-mediated functions like pollination and pest regulation, as well as refugia provision for emergency forage and wildlife shelter. By explicitly linking interventions to these spatial states, the framework provides an analytical foundation for evaluating how alternative strategies shape regenerative capacity in multifunctional landscapes.

To operationalize these concepts, the researchers developed a harmonized policy map by merging digitized land tenure segments from historical records with spatial layers from the Maasai Mara Wildlife Conservancies Association. This dataset was adjusted to approximate a snapshot of the tenure structure for the period 2000-2010, ensuring consistency with the era most relevant to the broader analysis. The map explicitly encodes historical tenure shifts, distinguishing areas that were historically organized as group ranches but have since transitioned to private ownership from those that have remained private. This categorization allows the model to capture path-dependent effects of past policy regimes, particularly the breakup of group ranches on current land configuration. By representing the spatial distribution of land tenure during this reference period, the dataset provides a basis for evaluating how legacy policy decisions continue to shape ecological and socio-economic outcomes across the landscape.

The water-pan dataset integrated georeferenced locations from OpenStreetMap with additional points collected through systematic field observations. Before combining these sources, the team performed a manual assessment via Google Earth to verify the existence of the water pans. The merged dataset provided a reliable baseline for assessing current water access across the landscape. To identify potential locations for additional water pans, the researchers applied a spatial multi-criteria siting framework restricted to areas with clay or silt soils and gently sloping terrain. Forested areas were excluded to avoid ecological disturbance, and a minimum distance of two kilometers from any existing water pan was enforced to prevent redundancy. A random spatial allocation procedure distributed 500 point locations across these suitable areas, incorporating considerations for community needs and livelihood diversification, including small-scale irrigation.

The spatial comparison of tenure scenarios utilized a regular grid to visualize where scenarios diverged from the baseline. The 90-meter predicted land-cover maps were summarized within approximately two-kilometer grid cells, with net change calculated as the difference in class share between the tenure scenario and the baseline. Cells with less than 50% of their area within Narok County were excluded from the analysis. The researchers reported percentage-point changes rather than relative percentage change because relative change can become unstable where the baseline proportion of a land-cover class is small. To assess whether the observed spatial patterns were sensitive to grid resolution, the analysis was repeated using a coarser five-kilometer grid, which produced consistent spatial patterns. This methodological choice ensures that the reported differences reflect substantive spatial reorganization rather than artifacts of grid scale.

The study’s limitations highlight that the spatial indicators diagnose structural conditions associated with regenerative capacity but do not directly measure ecological regeneration or livelihood outcomes. Greater forest connectivity or adjacency may create spatial conditions favorable to species movement and pollination, but the indicators do not establish that these functions actually occur or quantify their magnitude. Realized ecological responses depend on factors not represented by land-cover configuration alone, including vegetation composition, species-specific responses, and disturbance history. The authors emphasize that the results should be interpreted as identifying the relative potential of alternative interventions to create spatial conditions conducive to regenerative processes, rather than as direct predictions of ecosystem-service outcomes. This distinction is critical for sustainability assessment, as it underscores the need to interpret configuration-based indicators relative to underlying governance contexts when diagnosing regenerative capacity.

Subject of Research: Spatial indicators and landscape configuration in arid and semi-arid landscapes

Article Title: Designing regenerative landscapes: Using spatial indicators to diagnose landscape reorganization under alternative policies and management

Article References: Kotikot, S. M., Smithwick, E. A. H., Gergel, S. E., Nankaya, J., & Abila, R. (2026). Designing regenerative landscapes: Using spatial indicators to diagnose landscape reorganization under alternative policies and management. Environmental and Sustainability Indicators, 32, Article 101614. https://doi.org/10.1016/j.indic.2026.101614

Image Credits: AI Generated

DOI: 10.1016/j.indic.2026.101614

Keywords: landscape ecology, arid and semi-arid landscapes, sustainability indicators, land tenure, spatial modeling, Kenya, regenerative capacity, forest fragmentation, climate adaptation, ecosystem services, Designing, regenerative