Why we fail: stakeholders' perceptions of the social and ecological barriers to reforestation in southern Malawi

Reversing deforestation is a pressing challenge for planetary health and global development, yet the path to restoring forests on a globally significant scale is ill-defined, and government implementation of afforestation projects often generates conflict with user groups. Understanding the needs and values of forest-reliant stakeholders in developing countries and effectively engaging them in forest landscape restoration planning and implementation are critical steps toward meeting afforestation goals, and yet few examples of inclusive participatory processes applied in this context exist. Inspired by an approach used to mediate other natural resource conflicts, I structured a series of stakeholder focus groups in the Zomba-Machinga region of southern Malawi around Systems Thinking and Bayesian Belief Networks. Through these focus groups, I engaged forest-reliant households, academics, and government staff in a participatory process to explore the conditions of past restoration failure; identify the barriers to future reforestation success; and develop proposed solutions. Out of eight focus groups held, seven viewed poverty as the single greatest obstacle to afforestation success, and all seven perceived small business capacity building as one of the three most important factors in poverty alleviation, followed in order of frequency by improved agricultural practices (six of seven groups), and non-forest employment (three of seven groups). If each focus group could successfully implement the three factors they considered most important, the groups' perceptions that poverty could be alleviated ranged from a likelihood of 15.8% to 62.3%, with differences among focus groups that underscored inequities in social agency and vocational opportunities of the participants based on age and gender. This study illustrates where these diverse groups of stakeholders from a single region converge and diverge in their thinking, underscoring the importance of establishing durable participatory processes to work through conflicts, nurture trust, and collaboratively develop solutions that are relevant to local conditions in order for global forest landscape restoration to succeed.In order to analyze the data in Netica, a blank row will need to be added at the top of each data sheet, with the following script pasted in the first cell (A1): // ~->[CASE-1]->~ Funding provided by: University of British ColumbiaCrossref Funder Registry ID: http://dx.doi.org/10.13039/501100005247Award Number: Hampton Fund GrantPlease refer to the manuscript for a complete description of the methods for participant selection, data collection, and analysis. In brief, the data were collected in eight focus groups. Each group developed a Bayesian Belief Network, and then individual participants in the focus groups completed conditional probability tables for four of the nodes in the network. Networks were constrained to one primary node, three secondary nodes, and three tertiary nodes for each secondary node, totaling 13 nodes. Nodes were constrained to two states; one described a positive or desirable outcome, and one a negative or undesirable outcome. In each group, a conditional probability table was developed for the primary node based on the states of the secondary nodes, and for each of the secondary nodes based on the states of the adjoining tertiary nodes. The conditional probability table presented eight possible scenarios (in other words, combinations of secondary or tertiary node states that were read as a hypothetical scenario to the participants) in each case, so each focus group participant assigned probabilities of positive outcomes for the subject node to 32 scenarios for their group's Bayesian Belief Network. The data provided here were taken from those conditional probability tables, so that other researchers can reconstruct and analyze the Bayesian Belief Network for each group. In the data, individual participants are identified by a numeric code to protect their identities, and each row in the data represents one scenario, so there are 32 rows of data (four eight-scenario conditional probability tables) for each participant