12/6/20

Distribution pattern of tree cover relative to settlements

Tree cover increased with distance from settlements, shown by both area and percent tree cover per concentric circle. While the sum area of tree cover appears exponential for each concentric circle (Figure 1), when standardizing for the area of each concentric circle, a logarithmic pattern presents itself (Figure 2, Table 1).

Thus, it appears that while areas near settlements have little tree cover, tree cover increases steeply with increased radius and gradually flattens. This trend supports the explanation that lack of tree cover is related to the presence of people in the area as it indicates high deforestation where residents have most access with a diminishing slope as distance from home increases.

.

.

.

.

This explanation, however, assumes that tree cover was once evenly distributed across the landscape and that homes are equally likely to be built at any site, therefore not considering that homes may have been built in areas already without tree cover. Analyses considering temporal depth would be necessary to gain a better understanding of tree cover and settlement dynamics.

Because 6 of the 50 points had at least one concentric circle touching Lake Malawi, these numbers underestimate more as radius increases (Figure 3). It did appear from satellite imagery that, apart from proximity to the lake, most lack of tree cover near houses can be explained by agriculture.

Looking to the literature

According to interviews from a study taking place within the study area in Mwazisi, Malawi, the proximate causes of deforestation are agriculture expansion, tobacco growing, and brick burning, and the driving factors for these are population growth are “(a) population growth, (b) poverty, (c) expensive alternative building materials, (d) lack of awareness, (e) lack of resources, (f) lack of commitment from the tobacco companies, and (g) market system of the cash crops grown in the area” (Ngwira & Watanabe 2019).

Another study identified a tradeoff between time collecting fuel and expenditures; namely, that proximity to woody savannah or degraded forest leads households to buy fuelwood, but proximity to forest regeneration leads households to collect fuelwood by decreasing the necessary time spent collecting (Jagger & Perez-Heydrich 2016). This could lend a deeper understanding to the logarithmic trend in tree cover identified around settlements, where the balance of collection time and expenditures are mediated by the quality and distance of woody areas in the form of diminishing returns (Jagger & Perez-Heydrich 2016).

Future steps for a better analysis

After making the concentric circles around each built-up point, to account for protected areas, which would inflate tree cover, and water, which would deflate tree cover (particularly given high urbanity near Lake Malawi), complete the following steps:

1. Import or create polygons for protected areas and water
2. To remove polygons from concentric circles, use ‘difference’ at each radius, using ‘.map()’ to run the function over all points. Use this sample code for guidance
3. Calculate tree cover using ‘ui.Chart.image.byRegion()’. Calculate area of concentric circles minus protected areas and water using ‘.area()’
4. Calculate percentages by taking tree cover numbers from chart and dividing by each area

Sources:

Jagger, P. & C. Perez-Heydrich. 2016. Land use and household energy dynamics in Malawi. Environmental Research Letters 11(12). https://iopscience.iop.org/article/10.1088/1748-9326/11/12/125004
Ngwira, S. & T. Watanabe. 2019. An analysis of the causes of deforestation in Malawi: a case of Mwazisi. Land, MDPI, Open Access Journal 8(3):1-15. https://www.mdpi.com/2073-445X/8/3/48/pdf