University of Exeter logo

Research and Innovation blog

Home About Contact Toggle navigation Open menu

Data Science for Climate Resilience in East Africa

24 July 2020

6 minutes to read

Data Science for Climate Resilience in East Africa

About TIST

TIST is a network of farmers who have to date planted over 19 million trees. Started in Tanzania in 1998 by Bishop Simon Chiwanga and missionaries Ben and Vanessa Henneke, TIST has grown to over 90,000 members across 4 countries: Kenya, Tanzania, Uganda and India.

TIST is organised around the idea of a small group consisting of 6 to 12 participants who plant trees. In addition to the direct benefits of firewood, fodder, fruit and shade, trained farmers are employed by the programme to visit each farm and quantify tree growth, generating verified carbon credits which are sold on the international carbon market, with profits returned to the farmers.

The small groups follow best practices such as rotating leadership, Kujengana (“building each other up”), female participation, and servant leadership. The small groups are organised into larger ‘clusters’ which meet and share best practices for agroforestry. This network is another major benefit of TIST, which enables farmers to share information and provides opportunities for all members, especially women, to take on leadership positions.

Involvement with Exeter

The University of Exeter’s IDSAI, and Global Systems Institute are working with TIST to quantify and understand TIST’s impacts at landscape scales and on multiple sustainable development goals. We held a meeting at Exeter with farmers involved in TIST to outline a project which could help TIST’s growth and efficacy. Using TIST’s observations, remote sensing and fieldwork we are assessing carbon capture, regreening, vegetation resilience, soil health and biodiversity impacts. Additionally, thanks to a grant from the Turing Institute, we have also been able to analyse TIST’s organisational data to study how factors like female participation and engagement with cluster meetings affect tree planting.


TIST gathers a wealth of data on all aspects of their tree planting scheme, both social metrics such as the gender and age of participants, and physical metrics i.e. the age, quantity and species of trees on a farm. We analysed this data to identify overall trends for TIST groups in Kenya, Uganda, and India. We found that in all three countries the early participants of TIST tended to be a younger demographic and that longer established TIST groups tended to attend fewer cluster meetings than newer groups. In India, the ratio of women in TIST groups was significantly positively correlated with the number and diversity of trees planted by those groups. By creating a model of expected number of trees per Ha based on available TIST data for Kenya (as Kenya has the biggest TIST network), Figure 1, we identified a ‘successful’ geographical region where many groups had planted more trees than our models expectations, and also had a high diversity of tree species (Figure 2) and low levels of eucalyptus planting (which is highly destructive). Locating such regions allows TIST to explore factors that might help contribute to the region’s success and whether these lessons can be shared with others groups to improve their own outcomes.

Figure 1: Data vs model predictions for Kenyan TIST groups


Figure 2: Density and diversity of trees for Kenyan TIST groups with the identified ‘successful’ region circled.


With the use of satellite data we have monitored the effect that TIST has had on the groves of smallholder farmers in the vicinity of Mount Kenya, as well as across the landscape more broadly. We used 30m resolution data from the Landsat 7 satellite to detect greening trends from 2000 – 2019 within agricultural land. For this analysis we used Kendall Tau, a test which returns a value between -1 and 1, with negative values corresponding to a decreasing, or browning, trend and positive values corresponding to an increasing, or greening, trend of a pixel. With this we were able to detect the TIST effect; that is that over this time period, TIST groves became greener than the surrounding agricultural region, as seen in Figure 3. We also investigated the secondary effects that TIST is having beyond the boundaries of groves. Figure 4 shows that, on average, pixels located near to TIST groves display a greater greening trend than those which are further away, with a measurable TIST spillover effect present up to 200m away from groves. This spillover effect is due to the ecological benefits of agroforestry, such as reduced soil erosion and increased soil nutrient levels, in addition to social effects of having a nearby source of wood fuel, as TIST farmers are less likely to deplete surrounding woodland.

Figure 3: Distributions of pixel greening trends within TIST and non-TIST areas. Positive values represent pixels which have undergone greening.

Figure 4: Average NDVI Kendall Tau of TIST pixels and neighbouring pixels at an increasing distance.


TIST’s financial model depends on quantifying and verifying carbon sequestration. We assessed the effectiveness of TIST farmers in sequestering carbon dioxide in the trees planted across Kenya since the start of the programme. We analysed 4.7 million individual tree circumference measurements made by specially trained farmers between 2004 and 2018. We used empirical relationships known as allometric equations to convert each tree circumference to an estimate of how much carbon was stored in that tree, then scaled these up to estimate CO2 sequestration for each farm.

Figure 5: Carbon sequestration by TIST farmers in the Mount Kenya region since 2008. Darker green colours reflect more CO2 stored per grid cell (in units of tonnes per hectare).

The 53,000 Kenyan farmers in the TIST database had stored more than 1.2 million tonnes of CO2 in their trees as of 2018, with the Eastern slopes of Mount Kenya capturing the most (Fig. 5).

We also divided the data by dominant tree species to understand whether some species were particularly effective at sucking up carbon. Fig. 6 shows the average sequestration rates achieved by 5- to 7-year-old trees of commonly planted species. This suggests that native species such as acacia, fruit and nut trees like macadamia and popular fodder trees like grevillea may sequester carbon as well or better than the exotic eucalyptus species that TIST is seeking to discourage.

Figure 6: Average carbon sequestration rates per tree among medium-aged (5- to 7-year-old) trees of common species. Black bars reflect calculated 95% confidence intervals.


The TIST programme, as a farmer-led enterprise which spreads by word of mouth has been extraordinarily successful. Through planting over 19 million trees, TIST has helped some of India and East Africa’s poorest farmers improve their lives and livelihoods. A ‘bottom-up’ approach with an emphasis on attracting diverse participants and providing networking and leadership opportunities to all members has brought opportunity and empowerment to the regions where TIST is active.

Climate change is not a problem for people 20 years in the future. TIST is active in some of the regions that are experiencing the worst effects of climate change right now. At Exeter we hope to continue our collaboration with TIST by helping them to analyse and understand their organizational data and to understand the ancillary benefits of TIST beyond tree planting. We hope that we can help TIST become an international example of how local leadership and ecologically beneficial farming can lead to a brighter, greener future for everyone.

For more information please contact:

Dr Rudy Arthur
Lecturer in Data Science

Dr Tom Powell
Associate Research Fellow, Geography 


Back home