Natural vegetation cover and socioeconomic development of Caatinga

Background

      The relationship between forest cover and human socioeconomic development still is an open question as we have mixed evidences about it. In general, high forest cover is associated with poverty, but in many places, low forest cover is also associated with high poverty rates. The economic development of a forested region frequently happens through deforestation which generates economic growth, but if the gains are not equally distributed (i.e., through access to education, land, market), we often register loss of natural capital that are not translated to reduction in poverty and inequality over the long run. Furthermore, too much loss of natural capital might lead to a worsening of socioeconomic conditions, since the ES need to sustain well-being are lost. When the deforestation is limited to intermediate levels, it is likely that there will be space to build the basic infrastructure necessary for human socioeconomic development (e.g. food production, water and energy infrastructure, roads, schools, hospitals) without disrupting the ecological systems. In this paper we tested the hypothesis that gains in socioeconomic conditions over time will be optimized in places with intermediate amounts of natural capital. In this sense, we expected that municipalities with higher levels of forest cover will have lower improvement in social indices over time than the municipalities with intermediate levels of forest cover. While the municipalities with low levels of forest cover will have the same or lower socioeconomic conditions that municipalities with intermediate forest cover

General methods

      I gathered data on native vegetation cover (NVC) and land cover change from MapBiomas 4.0. The HDI, Gini Index, extreme povery and under five mortality data where gathered from Atlas Brasil Project. I used an algorithm called ClustGeo to group the municipalities according to their similarity in vegetation cover and deforestation rates into groups of deforestation stages for 1991. Then, I used the NVC criteria established by ClustGeo to classify the municipalities according to each category for the following years. Thus, depending on the municipal NVC in the years after 1991, it can be reclassified to another stage of deforestation. I conducted a repeated-measures ANOVA, using the deforestation stages as “treatment”. Then I did a pairwise comparison between groups within each year and between years

Descriptive results

      Between 1991 and 2018, the native vegetation cover for the entire Caatinga biome remained constant (mean = 50.43%; sd = 0.62), according to Mapbiomas data, with vegetation cover loss in some areas that were compensated by regeneration in other areas. The rates of deforestation for the whole Caatinga was 1.56% (sd = 0.51)

Figure 1. Native vegetation cover of municipalities in Caatinga in four different years

      The cluster algorithm allowed to split the municipalities into three deforestation stage groups, which we named Initial, for municipalities with NVC higher than 70%; Intermediate, with NVC between 40-70%; and Advanced, with NVC lower than 40%.

Figure 2. Dendrogram generated by ClustGeo evidencing the three deforestation stages in Caatinga. Blue rectangle represents the municipalities in initial deforestation stage; the yellow, municipalities at intermediate deforestation stage; and the red, the advanced deforestation muncipalities.

      The stage “Initial” group had 84.52% of native vegetation cover (NVC) on average (sd = 6.92), the stage “Intermediate” had 54.76% (sd = 11.07), and stage “Advanced” had 15.44% on average (sd = 9.01)

Figure 3. Distribution of native vegetation cover (%) in each deforestation stage in 1991. NVC - Native Vegetation Cover

      The deforestation rates for each group from 1986-1991, used to defined the deforestation stages in Caatinga. Deforestation stage on has 1.57% of deforestation rates on average (sd =0.89), the stage “Intermediate” had 2.81% (sd = 1.15), and stage “Advanced” had 2.28% on average (sd = 1.33)

Figure 4. Distribution of mean anual deforestation rates (Def. rate) between 1991 and 2010 in each deforestation stage.

Table 1. Repeated measures ANOVA for several socioeconomic outcomes

	HDI			Gini index			Extreme poverty			Under five mortality			HDA
	Chisq	Df	Pr(>Chisq)	Chisq	Df	Pr(>Chisq)	Chisq	Df	Pr(>Chisq)	Chisq	Df	Pr(>Chisq)	Chisq	Df	Pr(>Chisq)
def.stage	12.24297	2	0.0021952	12.32460	2	0.0021074	37.80203	2	0	14.17651	2	0.0008349	31.48637	2	0.0000001
year	99275.65166	2	0.0000000	618.55913	2	0.0000000	13728.93728	2	0	30148.97305	2	0.0000000	820.97228	9	0.0000000
def.stage:year	71.74041	4	0.0000000	27.45897	4	0.0000161	93.50772	4	0	46.30648	4	0.0000000	33.33707	18	0.0151893

      Over time, the gains in HDI were 109.09% in stage “Initial”, 102.42% in stage “Intermediate” and 97.10% in stage “Advanced” in mean. A pairwise comparison reveals that stage “Initial” had a gain of 6.67% more compared to stage “Intermediate” and 11.99% compared to stage “Advanced”. Also the stage “Intermediate” had a gain in HDI of 5.32% more than stage “Advanced”.

Figure 5. Evolution of Human development index (HDI) per deforestation stages. Letters are the contrast between each deforestation stage within the same year and across time.

      Over the 20 years time span of our analysis, there were mixed results in Gini Index. In stage “Initial” there were an increase of 2.40%, a reduction of 2.05% in stage “Intermediate” and an increase of 1.39% in stage “Advanced” in mean. A pairwise comparison reveals that stage “Initial” had an increase 4.45% bigger than stage “Intermediate” and 1.01% compared to stage “Advanced”. Also, the stage “Intermediate” had a 3.44% greater reduction when compared to stage “Advanced”.

Figure 6. Evolution of Gini index of income inequality per deforestation stages. Letters are the contrast between each deforestation stage within the same year and across time.

      There were consistent reduction in extreme poverty for all deforestation stages between 1991 and 2010. The stage “Initial” municipalities had a decrease of 55.16% in mean, while the stage “Intermediate” had 61.16% decrease and stage “Advanced” with 55.40%. A pairwise comparison reveals that the reduction in extreme poverty in stage “Initial” was 6.00% smaller when compared to stage “Intermediate” and 0.24% when compared to stage “Advanced”. The stage “Intermediate” had greater reduction in extreme poverty of 5.75% compared to stage “Advanced”.

Figure 7. Extreme poverty levels per deforestation stages. Letters are the contrast between each deforestation stage within the same year and across time.

      There were also great reduction in child’s mortality in all deforestation stages. all deforestation stages had a decrease around 70% in mean. A pairwise comparison revealed a not significant, but slightly bigger reduction in stage “Intermediate” (0.89% comparing to stage “Initial” and 0.30% comparing to stage “Advanced”). Between stages “Initial” an “Advanced”, stage “Advanced” had a greater reduction of 0.6%.

Figure 8. Under five mortality levels per deforestation stages. Letters are the contrast between each deforestation stage within the same year and across time.

      We also found a general reduction on height deficit by age in childs in all deforestation stages. In stage “Initial” the decrease between 2009 and 2018 aws 24.72%, in stage “Intermediate”, 21.63%, and 17.17% in stage “Advanced”. There were no significant differences between the deforestations stages in 2018, but in some years the municipalities in stage “Initial” had a statisticaly higher levels of HDA. We found a slightly bigger reduction in stage “Initial” (3.09% comparing to stage “Intermediate” and 7.56% comparing to stage “Advanced”). Between stages “Intermediate” an “Advanced”, stage “Intermediate” had a greater reduction of 4.47%.

Figure 9. Evolution of percentage of children with Height Deficit by Age (HDA) in 1210 municipalities within Caatinga Dry Forest. The horizontal lines are the means of each group and the vertical lines are the standard deviation in each year.

      Pairwise comparison between deforestation stages. Percentages are the differences for each socioeconomic and health variable. For HDI, gini, expov, u5mort the differences are for 2010. For HDA, the differences are for 2018.

…1	stage I-II	stage I-III	stage II-III
HDI	6.67% (z-score= -0.086; p-value= 1.0)	11.99% (z-score= 2.817; p-value= 0.1102)	5.31 (z-score= 3.374; p-value= 0.0211)
gini	4.45% (z-score= 4.431; p-value= 0.0003)	1.01% (z-score= 3.249; p-value= 0.0318)	-3.43% (z-score= -0.970; p-value= 0.9885)
expov	5.99% (z-score= 4.134; p-value= 0.0012)	0.24% (z-score= 2.674; p-value= 0.1571)	-5.75% (z-score= -0.975; p-value= 0.9882)
u5mort	0.89% (z-score= -0.599; p-value= 0.9996)	0.59% (z-score= -0.420; p-value= 1.0)	-0.29% (z-score= 0.126; p-value= 1.0)
HDA.dif	-3.09% (z-score= 1.112; p-value= 1.0)	-7.55% z-score= -0.248; p-value= 1.0)	-4.46% z-score= -1.444; p-value= 0.9999)