Understanding forest land conversion for agriculture in a developing country context: An application of the theory of planned behaviour among a cohort of Nigerian farmers

The global and Africa's loss of forest area was 4.74 and 3.94 million hectares between 2010 and 2020, respectively, placing Africa on top of the list on account of this loss across the world regions in the decade (FAO and UNEP 2020). Africa is losing its forest area progressively, whereas this loss is regressing globally (Ibid). Using datasets captured with the use of satellite images of the Landsat TM/ETM+, Arowolo and Deng (2018) reported that forest land decreased and increased by 30.1% and 19.5%, respectively, in Nigeria between 2000 and 2010, leading to a loss of 10.6% of forest land in the decade. Arowolo and Deng (2018) further reported that cultivated land reduced by 18.1% and increased by 34.2% between 2000 and 2010 in Nigeria, making the increased cultivated land to be 16.1% in the decade. Forest loss is on account of several factors, but boosting agricultural production is a major motive of decimating forests, which results in the loss of biodiversity (Badamasi et al. 2018; Fasona et al. 2020; Miyamoto 2020; FAO and UNEP 2020). Agricultural expansion is solely responsible for 70–80% of forest loss in Africa (Olorunfemi et al. 2021). This underscores the significance of forest land conversion for the purpose of agricultural expansion.

The importance of agricultural land area/use must not be underemphasised. It serves as the base of providing food, fibre, fodder and other necessities. It boosts livelihood opportunities and occupies 40% of the land cover (Hasan et al. 2020), making agricultural land a prime land use type. Especially in least developing countries like Nigeria, increasing human population puts the agricultural enterprise under pressure to expand (Olorunfemi et al. 2021). Nigeria is classified as the second of the nine countries that will account for 50% of the global population growth by 2050 (United Nations Population Division 2017). ‘The demand for food is resulting in inappropriate agricultural practices that drive large-scale conversion of forests to agricultural production’ (FAO and UNEP 2020). Hence, forest land areas are continuously converted. As opposed to the situation in Southeast Asia where agriculture for export drives land conversion, subsistence agriculture for the local market motivates this conversion in Africa (Doggart 2020).

Converting natural forest system to agricultural land impedes the capacity of the ecosystem to function, thereby hindering sustainability of the ecosystem (Buytaert et al. 2014; Fagerholm et al. 2016; Turner et al. 2016; Tolessa et al. 2017; Fu et al. 2017). Converting land for agriculture is a variant of deforestation because deforestation is essentially the transformation of forest land to non-forest land (Fasona et al. 2020; Doggart 2020). Forest degradation hampers biodiversity (Duraiappah et al. 2005; FAO and UNEP 2020; Fasona et al. 2020), and therefore weakens the adaptableness of the ecosystem (Corral-Verdugo 2009). More specifically, the functioning of forests as carbon sinks is negatively affected when lands are converted for agricultural use. Aside from energy use, forest land conversion is the largest contributor to CO₂ and other greenhouse gas emissions in Nigeria, considering that 60% of total emissions were on account of this conversion (Fasona et al. 2020). Greenhouse gas emissions herald into rising temperature and climate change (Olorunfemi et al. 2021). Reducing these emissions is central to mitigating climate change. Climate change is probably the greatest global environmental threat of our time, making international accords to resolve to mitigate the same by halting the global CO₂ emissions to net zero by 2050 (Ayompe et al. 2021). One-third of the greenhouse gas emission is on account of land use change cum agricultural activities in sub-Saharan Africa (Olorunfemi et al. 2021). The preservation of Africa's forests has been identified as a means of lightening the burden of climate change in the Reducing Emissions from Deforestation and Degradation (REDD) programmes (Olorunfemi et al. 2021).

Natural and forest-rich ecosystems optimally provide regulatory services that protect climate, prevent hazards such as landslides and wildfires, purify the atmosphere, preserve soil and detoxify water (Hasan et al. 2020). Farmers' forest land conversion for agricultural purposes, therefore, represents a dilemma warranting systematic examination. As opposed to this conversion, farmers have the choice of sparing forest lands for conservation purposes and cultivating high-yielding varieties of seeds. They also have the choice to regard forest land conversion as an indispensable route to food production. Whether or not they accelerate forest conversion, farmers also have the choice of embracing agro-forestry where food production and biodiversity could thrive in an integrated fashion (FAO and UNEP 2020). The point is that there are diverse courses of human action which have weighty implications on environmental health. This bares the relevance of models of human behaviour in understanding motivators of environmentally related actions. Natural resource disciplines are yet to optimally exploit the potentials of human behaviour models in understanding the dynamics of environmental resource use. Yet, man is in the anthropocene age, where actions in his ecological space have ushered in a new geological era (Lewis and Maslin 2015). Forest and environmental degradation are general samples of human behaviour. Effective conservation of environmental resources depends tremendously on human behaviour, thereby showcasing the huge relevance of human social research (St John et al. 2011; St John et al. 2013; Schultz 2011; Veríssimo 2013; Crandall et al. 2013; Nielsen et al. 2021; Travers et al. 2021; Chua et al. 2020; Häyrinen and Pynnönen 2020; Knapp et al. 2021; Cosyns et al. 2020; Crandall et al. 2018). It follows that human behaviour is central to achieving the much desired ecologically balanced environment. Hence, the employment of a prime human behaviour model, the theory of planned behaviour (TPB), in the examination of forest land conversion for agricultural purposes is called for.

The TPB is probably the most popular heuristic social science theory. It was developed by Ajzen and Fishbein (1980) and Ajzen (1985, 1991). The TPB is an extension of the theory of reasoned action which recognises individuals as rational beings and, therefore, upholds that they select their course of action after considering the favourability of the same (Ajzen 1985). According to the TPB, human behaviour is determined by behavioural intention. Intention is construed as the ‘indicators of how hard people are willing to try… to perform the behavior’ (Ajzen 1991). Intention to behave in a certain way is determined by a confluence of three factors – attitude, subjective norm and perceived behavioural control. ‘Attitude is defined as an individual's favorable or unfavorable evaluation of the behavior; subjective norm refers to the perceived social pressure towards the behavior; and perceived behavioral control is the personal assessment of the feasibility of executing the behavior in a given context’ (Yurieva et al. 2020). The TPB is very useful in creating behavioural interventions (Riebl et al. 2015). It has been employed in a wide variety of contexts, including the study of environmental behaviours (Yurieva et al. 2020). These environmental behaviours include farmers' tree planting (Zubair and Garforth 2006), water conservation (Lam 2006), usage of alternate transportation (Muñoz et al. 2016), recycling (Echegaray and Hansstein 2017), choice of tree species for reforestation (Osei et al. 2018), conservation of energy (Allen and Marquart-Pyatt 2018), reduction of flood risk (Allred and Gary 2019), low-carbon enhancement (Kaffashi and Shamsudin 2019), consumption of low carbon (Jiang et al. 2019) and agroforestry practices (Buyinza et al. 2020). In the context of this study, the implication of employing TPB is that forest conversion behaviour is determined by a partly novel construct – intention to convert forest land for agriculture. Further, this intention is collectively determined by three partly novel constructs – attitude towards converting forest land for agriculture, subjective norm of converting forest land for agriculture and perceived control of decision to convert forest land for agriculture. Therefore, the specific aims of this study included the prediction of forest conversion behaviour by the intention to convert forest. Another aim was the prediction of intention to convert by attitude towards, subjective norm of and perceived behavioural control of forest conversion. Analyses of items used in the measurement of variables were attempted, while the effects of gender, age and education on forest conversion behaviour were examined. These were accomplished among a group of Nigerian farmers.

The results of the univariate analyses of forest conversion behaviour are indications of the high extent of forest conversion among respondents in the study area. This is certainly contrary to the interest of optimal ecosystem functioning. Similar results were reported by Doggart et al. (2020) who conducted observations on 119 randomly selected plots from areas that have been deforested in Tanzania. Doggart et al. (2020) reported that 89% of plots were cultivated as opposed to 69% and 35% of plots that were used for grazing and charcoal production, respectively. Doggart et al. (2020) asserted that agriculture was the major driver of deforestation in Tanzania because there is no clear governmental policy banning the conversion of forest land for agriculture beyond protecting selected areas. Tesfahunegn, Ayuk and Adiku (2020) also reported that 84.3% and 63.3% of their respondents opined that arable land expansion was responsible for the state of forest degradation in northern and eastern Ghana, respectively. The high level of forest conversion in the current study vindicates the findings of Fasona et al. (2020) who indicated that agriculture land area increased drastically from 21% in 1986 to 37% by 2016 in in Oyo state, Nigeria. Indeed, the trend of forest land conversion for agriculture in the study area is inimical to environmental sustainability.

The higher extent of forest conversion behaviour among male respondents is supported by Meijer et al. (2016), who also found that men held stronger unhealthy attitude towards falling forest trees when compared with women. The higher tendency of forest conversion among men is probably born out of unequal land tenure rights bestowed to men and women. Customary laws are typically in favour of land ownership on the part of men as opposed to women (Marin and Kuriakose 2017). This situation could also be born out of greater connectedness and protectiveness to the environment among women (Lau et al. 2021). However, the current study shows that the differential means between gender subgroups were not significantly different. The literature is scant with regard to determinants of forest conversion behaviour, but related studies support the current finding: Tesfahunegn, Ayuk and Adiku (2020) reported that gender was not a significant factor affecting farmers' perceived severity of forest degradation in selected communities of eastern and northern Ghana. Soe and Yeo-Chang (2019) similarly reported that gender was not a significant determinant of willingness to participate in forest conservation in a forest-dependent community of South-Central Myanmar.

The finding of extreme age subgroups (16–30 years and 76 years and above) rather than the middling age categories (31–45, 46–60 and 61–75 years) as being significantly predisposed to lowered extent of forest conversion behaviour is quite interesting. It indicates that protectiveness of the forest space is a prerogative of younger and older persons, while middle-aged persons are significantly less protective. Tesfahunegn, Ayuk and Adiku (2020) similarly reported that age was a significant determinant of perceived severity of forest degradation among farmers in Ghana (p < 0.05). However, Soe and Yeo-Chang (2019) reported that age did not significantly influence willingness to participate in forest conservation (p > 0.05) in South-Central Myanmar. Ratsimbazafy et al. (2012) reported that older persons were more supportive and demonstrated greater confidence to participate in forest conservation in Madagascar, while Shan (2012) and Zhang et al. (2011) found that younger persons demonstrated greater willingness to participate in decision-making of urban green spaces and converting cultivated land to wetland in China, respectively. The literature is diversified with regard to the role of age in peoples' demonstration of forest protection.

Education had no significant effect on forest conversion behaviour. Soe and Yeo-Chang (2019) similarly found that education did not significantly influence willingness to participate in forest conservation. However, Tesfahunegn, Ayuk and Adiku (2020) asserted that education significantly affected perceived severity of forest degradation. The insignificantly inverse relationship between education and forest conversion indicates that increased education insignificantly predisposes to decreased forest conversion behaviour. This suggests the protectiveness of education against peoples' forest conversion behaviour.

Although the application of TPB in the examination of forest land conversion seems non-existent, other relevant studies have reported the performance of the constructs of TPB: Maleksaeidi and Keshavarz (2019) applied the TPB to study intention towards on-farm biodiversity conservation among 274 Iranian farmers. They found that the theory explained 53.4% of the variance in intention. Attitude towards biodiversity conservation was the best predictor (β = 0.59, p < 0.001), while subjective norm of biodiversity conservation was a better predictor (β = 0.40, p < 0.001). However, perceived behavioural control was not a statistically significant predictor. Empidi and Emang (2021) similarly engaged the TPB to explain behavioural intention of the public to take part in programmes aimed at protecting forested watershed areas in Cameron. They found that attitude (β = 0.432, p < 0.01) was the best predictor, while subjective norm was a better predictor (β = 0.190, p < 0.05). However, perceived behavioural control was not a significant predictor (β = 0.087, p > 0.05). The current findings are highly congruent with the reports of Maleksaeidi and Keshavarz (2019) as well as Empidi and Emang (2021) in terms of the relative strength of attitude, subjective norm and perceived behavioural control. Meanwhile, Deng et al. (2016) also used the TPB to examine ecological conservation behaviour among 1004 farmers of the Loess Plateau, China. Deng et al. (2016) reported that perceived behavioural control was the best predictor (standardised path coefficient = 0.50, p < 0.01), subjective norm was a better predictor (standardised path coefficient = 0.42, p < 0.01), while attitude was a good predictor (standardised path coefficient = 0.33, p < 0.01). In addition, behaviour was a positive determinant of intention (standardised path coefficient = 0.88, p < 0.01). In Uganda, Buyinza et al. (2020) examined the intention to adopt agroforestry among 400 famers and reported that attitude and perceived behavioural control were significant determinants of intention, but subjective norm was not. Contrary to the findings of Buyinza et al. (2020), Ofoegbu and Speranza (2017) found subjective norm to be the major determinant of intention to adopt sustainable forest use in South Africa. These indicate that perceived behavioural control and subjective norm are also strong determinants of intention depending on the context. Ajzen (1991) asserted that contexts and differential samples of behaviour affect the predictive ability of TPB constructs. The stronger the degree to which forest conversion is considered as advantageous (attitude), the more the acceptability of forest conversion in peoples' social network (subjective norm); the stronger the independent power to decide to convert forest (perceived behavioural control), the greater the resolve to convert forest (intention). The greater this resolve, the higher the extent to which people will engage in clearing out forest vegetation for agricultural purpose (behaviour).

There is an intensive extent of forest land conversion for agriculture, which is a cog in the wheel of optimal ecosystem functioning in the study area. Gender and education are trivial sociodemographic variables in matters of forest conversion. Meanwhile, middle age is predisposing to, whereas younger and older age is protective against greater extent of forest conversion. The more robust the attitude towards converting forest land for agriculture, subjective norm of converting forest land for agriculture and perceived control of decision to convert forest land for agriculture, the stronger the intention to convert forest land for agriculture. The stronger this intention, the higher the forest conversion behaviour. Socio-psychological variables of the TPB are apposite in instilling forest conservation. Socio-psychologically inspired interventions are, therefore, promising in ensuring that forests are handled with the consciousness of the need of future generations to fulfil their needs.