Impacts of spatio-temporal variability in forest Landscape metrics on soil quality, erosion severity and patch morphology in the Copperbelt Province of Zambia

Abstract

Studies focusing on the impact of forest and cropland fractal geometry (FRAC) on soil loss, erosion severity and patch morphology are rare and the gap is particularly large in the Miombo ecosystem leaving substantial uncertainty. Most studies have concentrated on land use land cover change (LULCC) and yet significant variations in soil properties have been shown by this study to be influenced by the geometric pattern of landscapes such as Forestlands. This means that soil nutrients, soil loss and erosion severity can be managed / manipulated by designing / adjusting the geometry and morphological characteristics of patches which presents a new cost effective and environmentally friendly approach to ecosystem management. Therefore, the link between LULCC and the intensity of fragmentation, soil loss and erosion severity is important. These findings are especially critical in highly heterogeneous landscapes such as the Miombo woodlands.

This study was carried out in the Copperbelt Province of Zambia. This predominantly mining province has undergone massive LULCC which has resulted in decline in the extent of natural ecosystems such as Forestlands / woodlands and Dambos leading to an increase in fragmentation. High rates of forest degradation and fragmentation have negatively affected the climate, soil properties, fauna and flora. The loss of forest cover also entails loss of plants and biodiversity that anchor the soil and protect it from soil nutrient loss. With increased soil erosion, agricultural production is expected to decline which may compel farmers to increase the use of expensive inorganic fertilizers or clear more forest in a bid to raise yields. This trend is impacting on global food security at the time when human population is increasing. It is therefore critical to quantify the past, present and future impacts of LULCC, forest fragmentation dynamics on soil quality, loss and erosion severity and the morphological characteristics of landscapes / patches for effective land use planning. However, this field of science has also remained largely unexplored in forestry and soil science leading to inadequate information necessary to trigger effective mitigation action resulting in negative effects on ecosystem function.

This study aimed to quantify the impacts of spatio-temporal variability of LULCC, landscape metrics on soil quality, loss and erosion severity and on morphological / shape characteristics of patches. This was done by assessing the LULCC from 1984 and predicting the same to 2050, evaluating the impact of the Fractal Dimension (FRAC) on selected soil parameters, assessing the impact of FRAC on soil loss and erosion severity and modelling the morphological and shape descriptors of landscape / patches for use as indicators in spatial planning for sustainable management of forest and soils.

The first study analysed the LULCC between 1984 and 2016, and projected to 2050 because the pattern of Miombo woodland conversion to other land uses and the attendant impacts on vital Miombo ecosystems such as Dambos is not well understood. Land use land cover (LU/LC) maps were produced through supervised image classification in ArcGIS 10.3 and ENVI 4.7 and used as input in subsequent studies (Chapter 3-5). The effects of LULCC on the extent of Miombo woodlands and Dambos was determined by intersecting layers of Croplands, Settlements, Plantations, Grasslands and Barelands on Woodland and Dambo pixels. Prediction of future LULCC was done using the land change modeller (LCM) in TerrSet. Woodlands and Dambos were predicted to decrease by 26.4% and 2.0%, respectively, by 2050. Conversion to Cropland was the highest contributor to the loss of Woodlands and Dambos accounting for more than 54% of the total loss. Expansion of Croplands caused a decline in Woodlands and Dambos. Therefore, it was recommended that sustainable agriculture practices that do not rely upon land expansion should be adopted. The second study quantified the extent of fragmentation and the impact of FRAC on some soil parameters using LU/LC of the Copperbelt Province of Zambia from the first study. The LU/LC were processed into persistent forest and croplands (1984 to 2019). FRAC and other landscape metrics were computed in FRAGSTATS 4.2. The impact of FRAC on soil nutrients was assessed in the persistent forest layer (1984 to 2019) which was divided into FRAC-1 (Areas of Frac index ≤1.25 with simple shapes of patches) and FRAC-2 (areas of Frac index > 1.25 to 1.5, with more open, complex shaped patches); in sandy and loamy soil. The persistent layers were used in order to ensure that the only factor causing the variation in soil properties was the forest / cropland geometric pattern. Soil samples were collected for analysis of soil texture, pH, phosphorus (P), nitrogen (N), ammonium-nitrogen (NH4-N), nitrate-nitrogen (NO3-N), carbon (C) and potassium (K) concentrations. In areas with simple shapes of patches (FRAC-1), N, NH4-N, NO3-N, C and K concentrations were higher compared to areas with complex shaped patches (FRAC-2) in loamy soils. Forest geometric pattern affected the levels of soil parameters, therefore, incorporating fragmentation analysis in forest and soil conservation is recommended.

The third study compared the extent of soil loss and determined the impact of FRAC of persistent Forests and Croplands (1984-2019) on erosion severity in sandy and loamy soils. The Revised Universal Soil Loss Equation (RUSLE) was applied in ArcGIS 10.3. The area covered by the Low Erosion Severity class (<5 t.ha-1yr-1) reduced by a net change of 15.05% while the Moderate Erosion Severity (5 – 12 t.ha-1yr-1) and High Severity class (>12 t.ha-1yr-1) classes increased between 1984 and 2019. The mean soil loss (MSL) varied according to differences in the geometric pattern and soil textural classes. Areas with simple shaped patches had higher MSL by 10.2% and 18% in Croplands and Forestlands, respectively, than in areas with complex shaped patches in loamy soils. Overall, the MSL in Croplands was 95% higher than in Forestlands considering all fixed factors. Cropland and Forestland geometric pattern affected soil loss and erosion severity, therefore, it is suggested that fragmentation analysis should be incorporated into land use planning for sustainable agriculture, soil, and forest management.

The fourth study modelled the morphological spatial pattern and shape characteristics of persistent Forest and Cropland areas (1984 to 2019). Morphological Spatial Pattern Analysis (MSPA) used GUIDOS 3.0 to compute the Cores, Islets, Loops, Bridges, Perforations, Edges, Branches and Contortion while shape descriptors (Roundness, Circularity, Convexity, Elongation, Width and Length, Number of Holes and Directionality) were computed in ImageJ1.53 software. The study found that forestlands recorded the highest number of Core areas while Croplands had the highest amount of Islets with over 63% in both FRAC metrics. The implication of the MSPA results is that landscape connectivity is reduced by a decline in Core areas in the forest class, which blocks the migration of species, affects ecosystem integrity, nutrient flows and impacts on the provision of ecosystem services.

Forest areas (FRAC-1 and FRAC-2) had lower average values of Elongation and Convexity index than Croplands. Forest areas also recorded higher Contortion index than Croplands suggesting more complex patch shapes in natural landscapes. We identified the dimensional characteristics of FRAC areas for easy application in landscape design and restoration programmes. For instance, FRAC-1 forest areas would be restored by ensuring an average patch length and width of 800 m x 588 m respectively; with a maximum contortion index of 188 while for FRAC-2 Forestlands, the mean length and width would be about 28,818 m x 20,169 m with contortion index of over 2,000 corners. This indicates high shape complexity in FRAC 2 areas. This study concluded that the spatial and temporal morphology of natural landscapes (Forestlands) differed significantly with human induced land uses (Croplands) with implications on ecological connectivity and the provision of ecosystem services in general. It is recommended that strategies like land zoning, practicing safe and climate smart agriculture, effective enforcement of laws and policies, strengthening institutions and developing mechanisms for prescribing forest/vegetation geometry patterns should be implemented.

The overall conclusion of the study was that the geometric pattern of Forestlands had significant impact on soil quality, loss and erosion severity. This means that soil quality, loss and erosion severity can be managed by manipulating the geometric and morphological characteristics of patches / landscapes which presents a new approach to ecosystem management. Years of uncoordinated and uncontrolled land use activities will reduce the area of persistent forests and if the situation is not addressed, significant amounts of topsoil / nutrients and vegetation would be lost. The loss of nutrients will lead to a decline in agricultural production which will affect the already unstable food security situation in the area. Therefore, there is urgent need to incorporate FRAC analysis in land use planning.