Evaluating the environmental impacts of large-scale afforestation in Pakistan
The Billion Tree Tsunami Afforestation Programme in Pakistan helped increase tree cover in the Khyber Pakhtunkhwa province. However, satellite and administrative data reveal that it had only limited short-term effects on rainfall, temperature, and air quality. Additional monitoring and data are needed to assess the direct and indirect environmental impacts of afforestation efforts in the long run.
Many countries are making substantial investments in large-scale afforestation – planting trees on land without existing forest cover as a countermeasure to deforestation – as part of their climate mitigation and adaptation strategies.
Notable examples include India’s tree-planting pledges (2017), Mexico's Sembrando Vida (Sowing Life) programme (2019), Kazakhstan’s two billion trees project (2020), Turkey's ‘Breath for the Future‘ (2021), and Mongolia’s 1 Billion Trees Project (2021). Private firms, including major oil and gas producers, are also turning to tree planting as a way to achieve net-zero emissions.
Despite the growing popularity of such initiatives, there is limited rigorous evidence of their effectiveness in improving livelihoods and restoring ecosystems. Ecologists caution that poorly designed afforestation efforts can harm local environments, disrupt agriculture, strain water resources, and displace communities.
However, when implemented thoughtfully (for example, by selecting appropriate tree species and planting them in ecologically suitable areas), these programmes have the potential to sequester carbon, regenerate forests, and deliver essential ecosystem services such as flood mitigation.
Measuring the impact of Pakistan’s flagship climate initiative
Against this backdrop, Pakistan launched the Billion Tree Tsunami Afforestation Programme (BTTAP) as a flagship climate initiative. Implemented between 2014 and 2017 in the Khyber Pakhtunkhwa (KP) province, the programme aimed to plant or support the natural regeneration of one billion trees across public, communal, and private lands by establishing protected enclosures. Figure 1 shows the spatial spread of enclosures (for natural regeneration) and block plantations established by BTTAP.
Figure 1: Map of Khyber Pakhtunkhwa with enclosures and block plantations
We study the environmental impact of BTTAP in Pakistan using a combination of satellite imagery and administrative data. We begin by documenting the gain and loss of tree cover in the areas where the programme was implemented.
Next, within these areas, we assess the programme's impact across sub-units that were more versus less exposed to BTTAP activities on three environmental outcomes: precipitation, temperature, and air pollution. In addition to measuring these direct impacts, we investigate whether the programme has ‘spillover’ effects on downwind areas that did not directly host BTTAP, as trees can influence environmental conditions beyond their immediate location.
Did tree cover in Khyber Pakhtunkhwa increase?
We document a net tree cover increase of about 300 km2 in BTTAP areas from the start of the programme until 2020 (see Figure 2), which amounts to 6% of the area in which the programme was implemented. In contrast, the rest of KP experienced negligible changes in forest cover over the same period.
It is important to note that areas with and without exposure to BTTAP have important underlying differences. Since a large part of the programme consisted of enclosures around existing forests, sites where BTTAP was implemented already had six times more tree cover at baseline than non-BTTAP locations. BTTAP locations also had higher baseline rates of deforestation, which decreased significantly once the programme was rolled out. This is likely due to the establishment of enclosures and increased monitoring efforts. Thus, BTTAP potentially increased net tree cover by both increasing tree growth and reducing deforestation.
Figure 2: Changes in tree cover relative to 2000
How did BTTAP affect rainfall, temperature, and pollution?
We evaluate the impact of BTTAP using an event study at the Union Council (UC) level – the smallest administrative unit for which we have spatial boundaries. Rather than classifying areas as simply treated or not treated, we assess programme intensity by the share of UC land covered by BTTAP. To account for pre-existing trends, we control for time-variant and time-invariant effects on the yearly and district levels. We validate our approach by checking for similar pre-programme trends in BTTAP and non-BTTAP areas. To capture spillover effects, we use data on wind speed and direction to measure downwind exposure.
We find that a 10 percentage point increase in BTTAP exposure in a UC raised average rainfall by 0.5-0.8% between 2017 and 2019 (Figure 3).
Downwind, while there are no statistically significant impacts overall, we observe a small but statistically significant increase in rainfall in 2018. Impacts on rainfall in other post-intervention years are quantitatively similar but statistically insignificant (Figure 4). These effects, along with the direct rainfall impacts, imply that the programme may have had a modest positive effect on rainfall.
Figure 4: Effects of downwind exposure on (log) precipitation
Pre-treatment trends in temperature were similar across different levels of exposure to BTTAP. However, no significant changes in minimum or maximum temperatures were observed after the launch of BTTAP in treated UCs, either directly or in downwind areas.
The results for air pollution are inconclusive due to pre-existing differences between pollution levels in treatment and control areas. In areas with a higher downwind exposure to the programme, PM 2.5 levels were initially higher but lowered after BTTAP implementation, suggesting a potential small reduction in air pollution. However, this change cannot be definitively attributed to the programme due to the presence of pre-trends.
Looking ahead: Long-term impacts of BTTAP
The findings suggest that BTTAP led to notable increases in tree cover, but the direct and indirect environmental impacts remain uncertain or limited in the short term. Overall, while the programme may have had small positive spillover effects on rainfall, these effects are not consistently large or statistically robust.
However, it is possible that these effects take longer to materialise and may become more pronounced over time. This would depend on whether the trees planted or restored through the programme continue to grow in the years ahead. As satellite data currently extends only up to 2020, updating this analysis with additional years of data will be valuable for assessing longer-term impacts.