Pandemic-induced shocks and shifts in forest-based livelihood strategies: learning from COVID-19 in the Bia West District of Ghana

The COVID-19 pandemic has reshaped societies and will continue to do so. Despite its salience, micro-scale evidence on how this pandemic reshapes the livelihood strategies of forest communities in sub-Saharan Africa are lacking. To bridge this lacuna, this paper analyses the dynamics around forest-based livelihood strategies in the context of the COVID-19 pandemic in the Bia West District of Ghana. Key informant interviews (n = 8) and a survey of forest-dependent households (n = 100) were conducted to generate relevant data. The study identified fuelwood harvesting, medicinal plants extraction, fruit-gathering and beekeeping as the four livelihood activities that were predominantly practised in the study communities during the COVID-19 pandemic. The analysis established an increase in the number of households that diversified into fuelwood harvesting. Regarding medicinal plants collection as a livelihood strategy, less than 10% of households either diversified or intensified this practice with similar charges recorded in fruit-gathering and beekeeping. The logistic regression disclosed gender, household size, education and income, as the socioeconomic variables that significantly predict livelihood diversification and intensification during the COVID-19 pandemic. Additionally, the financial, social and physical asset base of households significantly shaped livelihood diversification. Diversification into fuelwood collection, medicinal plants extraction and beekeeping were the strategies that showed a significant positive correlation with the well-being outcomes of forest-based households. While this paper provides fresh evidence to inform the vulnerability dimension of the sustainable livelihoods framework, it further calls for policy interventions to build pandemic-resilient livelihood strategies around forest communities.


Introduction
Forest landscapes around the world are subjected to rapid transformation (Garcia et al 2020, Kimengsi et al 2020, 2022. For instance, annual forest cover loss in the global south stands at 12 million ha (Weisse andGoldman 2017, Garcia et al 2020). These rapid transformations potentially signal the rise in future health challenges, including pandemics. It has therefore become imperative to pay more attention to the links between forestry and human health (Tollefson 2020, Saxena et al 2021. Furthermore, the literature on forests and rural development has established the fact that the livelihoods of resource-dependent communities are reshaped during shocks (Shackleton et al 2011;Pretzsch et al 2014). In this regard, the sustainable livelihoods framework is a useful analytical lens to examine livelihood dynamics across marginalized natural resource-dependent communities that are vulnerable to shocks (Ellis 2000, Scoones 2009). Some of these shocks sufficiently documented in livelihood studies include political instability, macroeconomic shocks (e.g. a fall in price, inflation), and environmental shocks (Paumgarten and Shackleton 2011, Coulibaly 2015. However, issues linked to the dynamics around natural resource-based livelihood systems during shocks need to be continuously studied, especially in light of the COVID-19 pandemic in sub-Saharan Africa (SSA).
Studies in the global south hold that households diversify into forest and environmental resource extraction during health shocks (Hunter et al 2007, Saxena et al 2021. Such increases in diversification are related to mass in-migration and the increase in reliance on forest resources including fuelwood (Mbiba et al 2019, Saxena et al 2021. Such reliance could trigger improvements in well-being. For instance, empirical evidence from Ghana, Uganda, Rwanda and Tanzania also predict that wild food and other forest-based products contributed positively to household well-being outcomes including nutritional needs (Cooper et al 2018). Local households in the Brazilian Amazon that diversified into extracting resources during climate and health-related shocks also reported an improvement in their well-being (Pattanayak and Sills 1998). However, the extent to which households witnessed changes in well-being as a result of shifts in forest-based livelihood strategies remains relatively less understood in SSA.
Sub-Sahara Africa hosts about 15% of the total global forest cover, with over 80% of its population depending on the forest for their livelihoods (Adongo et al 2019, FAO 2020. This sub-region is not spared from the grievous impact of the COVID-19 pandemic. For example, it is reported that in Ghana, the COVID-19 pandemic caused over 1400 deaths with most of these deceased individuals being household income earners (GHS 2022). To curtail the spread of the virus, a stringent lockdown of all social gatherings including access to markets and trade was introduced. These measures further affected the livelihoods of many households (Aduhene and Osei-assibey 2021).
From the macroeconomic perspective, COVID-19 contributed to the fall in the annual growth rate of Ghana's Gross Domestic Product from 6.8% to 1.5%-the least annual growth rate since 1983 (MOF 2020). This has manifested in high rates of unemployment and severe economic hardship (Aduhene and Osei-assibey 2021). While macrolevel evidence exists in Ghana, micro-level analysis of livelihood response around natural resource sites (e.g. forests) is lacking (Bukari et al 2021). Therefore, it is unclear whether during the pandemic forest communities engaged more in diversification, and/or intensification in most contexts of SSA (including Ghana). Using evidence from the Bia West District, this paper: (a) investigates the changing livelihood strategies of forest-based households during the COVID-19 pandemic, (b) analyses the determinants of changing forest-based livelihoods strategies during the pandemic and (c) analyses the well-being implications of the changing forest-based livelihood strategies. With growing threats that future pandemics could be expected, this evidence is relevant to inform Ghana's forest management policy. Furthermore, despite having experienced earlier pandemics, attempts to develop pandemic-resilient livelihood strategies for forest-dependent communities are lacking in Ghana, as in many parts of SSA. This further exposes forest-based livelihood systems to future vulnerabilities (Chakraborty andMaity 2020, Lovejoy 2021).

Forest-based livelihoods in Ghana
Households engage in various forms of strategies to secure a living (Scoones 2009). With the onset of the COVID-19 pandemic, there is renewed evidence to explore the shocks and vulnerability dimensions of the sustainable livelihoods framework with a focus on forest-based communities (figure 1) in sub-Sahara Africa (Ayanlade andRadeny 2020, Kimengsi andMairomi 2021). Responding to these susceptibilities, an extensive body of knowledge affirms that households either diversify or intensify their livelihood strategy. Diversification in this study relates to the incidence whereby a household engages in a new forest-based activity to either support consumption, cash income or both (Wunder et al 2014). With intensification, a household tends to engage more often in a livelihood activity that was practised before the shock (Debela et al 2012). Many studies report that the choice of a livelihood diversification or intensification strategy is a reflection of household assets and socioeconomic characteristics (Stakhanov 2010, Debela et al 2012.
For instance, it is established that poor households, usually with large family sizes and less education, often diversify into forest-based activities more often than they intensify (Illukpitiya andYanagida 2008, Campbell andLuckert 2002). Empirical studies from South Africa affirm that many poor households diversified into forest and environmental resource extraction during health shocks which manifests in ailment or the death of a breadwinner of the household (Hunter et al 2007, Paumgarten andShackleton 2011). Recently, studies have categorized livelihood strategies into forest-based and non-forest-based strategies (Kimengsi et al 2020, Nerfa et al 2020. In the case of the Bia West District of Ghana, the applicable forest-based livelihood strategies include fuelwood collection, medicinal plants extraction, cocoa agroforestry, beekeeping, fruit-gathering, timber harvesting and livestock rearing.

Study area
Ghana is situated between latitudes 4 • and 11.5 • north of the equator with a population of 30 792 608 and a land area of 238 535 km 2 (FAO 2020, GSS 2021). Western north-the study region is one of the 16 administrative regions within the high vegetation zone of Ghana with a population of 880 921 (GSS 2021) and about 1108 COVID-19 recorded cases.
Bia West, the study district (figure 2) has a biosphere reserve which comprises a continuous block of two adjacent forest reserves, namely the Bia Resource Reserve and the Bia National Park (Asamoah 1999). These reserves of biodiversity importance are being encroached upon through illegal logging and agricultural expansion (Asare et al 2014). The study communities, Gyesewobre and Driver-Krom, are two of the 55 forest-based communities in the study district that were randomly selected for the study (figure 2). Cocoa agroforestry, practised harmoniously with the surrounding primary/secondary forest is the major income activity for more than 60% of households in these communities (Ashiagbor et al 2020). The key forest resources extracted before and during the COVID-19 pandemic are shown in table 1. These resources were mostly extracted from the buffer zone and on farms before COVID-19. However, some households extended these activities into the forest reserves during the COVID-19 pandemic, as recounted during field investigations.

Data collection
This study draws from two research instruments (appendix I); a semi-structured questionnaire and a key informant interview guide. Besides capturing questions on households' socio-economic conditions, the questionnaire also captured the livelihood strategies adopted prior to the pandemic, their changing nature and well-being implications. Furthermore, questions linked to the determinants of changing forest-based livelihood strategies and their implications for forest management were captured. These instruments were designed and validated between February and April 2021. Three foresters in the study communities were selected as research assistants and trained for three days prior to data collection. Data for this study were collected between April and June 2021. In the first phase of data collection, a total of eight key informant interviews on (a) how households perceived COVID-19, (b) the livelihood strategies practised before and during COVID-19 and (c) how these livelihood activities were affected during the COVID-19 pandemic were conducted. Semi-structured household interviews constituted the second phase of data collection. The research instruments were pre-tested (n = 8) at Asuopiri, one of the forest fringe communities within the study district, and further refined to capture relevant aspects of forest-based livelihood activities.
To ensure representativeness, a total of 104 households were randomly selected using the Yamane (1967)    represent the sample size, N is the total population and e is the margin of error. It should be noted that four household-heads, two from each community, did not make themselves available for the interview. This reduced the sample of the study to 100 households. Data was mostly collected in the evening when most household members have returned from their farms. A recall period of one year (six months before and during the pandemic) was established to minimise data inconsistency (Kimengsi et al 2020, Nguyen et al 2020.

Data treatment and analysis
Qualitative data obtained through key informant interviews were transcribed and analysed thematically. Codes and themes were assigned to the data and subsequently elaborated in narratives such as the key livelihood activities practised in the community before and during the COVID-19 pandemic, and their well-being implications. Additionally, data regarding income, food availability, savings and health were also explored. The quantitative data obtained through semi-structured households interviews were coded in the Statistical Package for Social Sciences and transferred into the R software (version 4.0.1) (R Core Team 2020) for both descriptive and inferential statistical analysis (table 2).

Estimation strategy
The econometric analyses focused on the relationship between socio-economic and asset determinants of diversification and intensification of forest-based livelihood strategies during the COVID-19 pandemic. The response variables are binary with 1  denoting the incidence whereby a household diversified into or intensified a forest-based activity during the COVID-19 pandemic, and 0 otherwise. The analysis is based on the underlying assumption that the choice of a livelihood strategy is a reflection of households' assets and socio-economic characteristics (Iiyama et al 2008). The econometric model for estimating this relationship (logistic regression) is as specified: where Log (odds) in the model denotes the conditional probability of an event occurring to the ratio of the event not occurring; α is the estimated response variable when the independent variables are held constant; β 1 and β n are the slope coefficients of the first and last regressors; X 1 and X n are the first and last regressors; and ε is the stochastic error term controlling for possible antecedent and extraneous factors that may have influenced the model.
It is worth mentioning that none of the households interviewed diversified into or intensified the practice of cocoa-agroforestry although it is the major economic activity in the study areas. Hence, the analysis focused more on other livelihood activities since results from the qualitative interview suggest that most households diversified into four key forest-based activities (fuelwood collection, medicinal plants extraction, beekeeping and fruit-gathering). Additionally, field realities informed by qualitative interviews revealed that diversification into forest-based activities was more common in the study area than intensification during the COVID-19 pandemic. Given this, the analysis focused more on diversification. A total of eight models: four models each were built to estimate the socio-economic and asset determinants of diversification using the four key livelihood strategies identified.
One model for each was built to estimate the socio-economic and asset determinants of a household intensifying in at least one of the four livelihood activities during the COVID-19 pandemic. Predictors with a variance inflation factor below the threshold of five (see supplementary table 1 (available online at stacks.iop.org/ERL/17/064033/mmedia)) were used since this indicates no multicollinearity between the predictors (Akinwande et al 2015). Also, models with a p-value <0.05 and a high Chi-square value were considered fit by using the likelihood ratio test as a determining criterion (Moreira 2003). The p-values of the logistic regression estimates were corrected for multiple comparisons to avoid the estimates being significant by chance (Gelman et al 2012). Moreover, the pseudo R 2 was used to determine the predictive power of the model (R 2 > 0.2 indicates a model with good predictive power) (Mcfadden and Zarembka 1974).
An equally weighted composite index of wellbeing outcomes was constituted comprising income, food availability, health, and savings (Zill 2006) as shown in equation (2).
where W i is the composite index score measuring well-being, Σ n i=1 X i is the summation of the first wellbeing outcome to the last well-being outcome, and N is the total number of the well-being outcome. This informed the well-being implication of the key livelihood strategies practised during the COVID-19 pandemic using the ordinary linear regression (equation (3)) as summarized below: A i and A n indicate the first and the last forest-based livelihood strategies. This is based on the supposition that the livelihood activities of forest-based households contribute to well-being outcomes during shocks (Harbi et al 2018). The point-biserial correlation was used to assess if a relationship exists between the forest-based activities practised and well-being during the COVID-19 pandemic.

Results
The results are presented following the three specific objectives earlier highlighted as follows: (a) an investigation of the changing livelihood strategies of forest-based households during the COVID-19 pandemic, (b) the analysis of the determinants of changing forest-based livelihoods strategies during the COVID-19 pandemic and (c) the analysis of wellbeing implications of forest-based livelihood diversification and intensification strategies during the COVID-19 pandemic.

Socioeconomic characteristics of households
Supplementary table 2 shows that the majority of household heads in the study area are males (76% vs. 24% females). Most of the households fell within the age category of 23-60 years (92%) and the rest are 61 years and above. Most of these household heads are married (71%); while the rest are single (11%), divorced (10%) and widowed (8%). About 52% of the households consist of 1-10 members (small family size) while 48% have a family size of 11 and above. Some household heads (34%) have no formal education, 35% have primary education, 22% secondary education with 4% and 5% having vocational/technical and tertiary education respectively. Close to 56% of the households have a small land size (1-5 acres), 29% hold lands between 6 and 10 acres while 15% have a large size of land (above 11 acres). Most (67.4%) of the households engaged in cocoa farming. This is followed by trade (11.4%), and livestock rearing (9.9%). The rest of the households are civil servants (7.8%) and small-scale miners (3.5%).
Most households (63%) either diversified into or intensified these activities for subsistence. Some 11% either diversified or intensified for cash income or to enhance savings, while 23% diversified or intensified for both subsistence and cash income.

Shifts in forest-based livelihood strategies during the COVID-19 pandemic
As earlier recounted, the pandemic induced shifts in livelihood strategies; some households diversified or intensified, while others did not. About 23.2% of households were involved in fuelwood harvesting before the COVID-19 pandemic. This subsequently increased to about 25.6% during the pandemic ( figure 3). Out of the 25.6% of the total households who engaged in fuelwood collection during the pandemic, 10.3% represent households who diversified into other activities. About8.2% intensified while 7.0% reported no change. Medicinal plants extraction increased from the pre-pandemic figure of 15.2% to 18.9% during the pandemic. Out of the total of 18.9% of households engaged in this practice, about 7.5% diversified, 6.1% intensified and 5.3% did not. Close to 18.5% of the households engaged in fruitgathering before the pandemic and 19% during the pandemic. Out of the total of 19% of households who engaged in fruit-gathering during the COVID-19 pandemic, 7.5% diversified, 6.1% intensified, while 5.3% opted for business-as-usual. Lastly, households who engaged in beekeeping increased from 2.2% before the pandemic to 3.6% during the pandemic. Out of the total of 3.6% households, 1.4% diversified, 1.2% rather intensified and 1% did not change their activities. Table 3 indicates that male-headed and educated households were 2.268 and 5.577 times less likely for male-headed and educated households to diversify into fuelwood collection than their female and uneducated counterparts respectively. That is, female-headed households and uneducated households were more likely to diversify in fuelwood collection during the COVID-19 pandemic. Households with members above 6, and those with dependents above 8 were 1.315 and 2.769 times more likely to diversify into fuelwood harvesting respectively   Concerning beekeeping, the result showed that educated households, migrants and households who are members of community organization(s) were 1.650, 1.831 and 1.146 times more likely to diversify into this livelihood activity than their counterparts respectively. Households who had access to credit during the COVID-19 pandemic were 1.752 times less likely to engage in fruit-gathering than their counterparts who did not. On intensification, the results highlighted that male-headed households were 1.531 times less likely to intensify the practice of at least one of these strategies than their female counterparts. Also, migrants on the other hand were 1.993 times more likely to intensify the practice of at least one of these strategies.

Households' asset determinants and shifts in livelihood strategies during the COVID-19 pandemic
Physical assets, mirrored through mobile phone ownership, financial asset (access to credit) and social assets (organizational membership) predicted the likelihood of households diversifying into forestbased activities during the COVID-19 pandemic (table 4). For instance, households with mobile phones diversified into medicinal plants collection and beekeeping by 1.306 and 1.368 times respectively more than their counterparts who do not have mobile phones. Households with financial assets (access to credit) were less likely to diversify into fuelwood by 1.895 times than their counterparts with no access to financial credit. Similarly, households who are members of community organizations stood a greater chance of diversifying by 1.641 times into beekeeping than their counterparts who do not belong to any community organization. On livelihood intensification, the results showed that households who own land (natural assets) were more likely (1.863 times) to intensify in at least one of the livelihood strategies than their counterparts who do not. Additionally, those with financial assets (savings) were 2.573 times less likely to intensify in at least one of the strategies.

Well-being outcomes of changing livelihood strategies during the COVID-19 pandemic
The well-being outcomes (changes in income, food availability, health, and savings) as a function of shifts in livelihood strategies show that households who diversified into fuelwood had about 0.969 times significant additional improvement in their wellbeing than their counterparts who did not diversify into fuelwood (table 5). In other words, while the well-being of diversified households increased, an additional increase of 0.969 points was reported for fuelwood diversifying households. Ceteris paribus, diversification into fuelwood harvesting contributed around 5% to the well-being of forest-based households. Households who diversified into medicinal plants extraction also recorded 0.681 times additional improvement in their well-being than their counterparts who did not diversify into medicinal plants. Diversification into medicinal plants contributed about 13.4% to the well-being of households while households who diversified in beekeeping also recorded a 3.1% improvement in their well-being. Furthermore, households who intensified medicinal plants extraction recorded about 0.614 times more improvement in their well-being than those who did not. Also, medicinal plants intensification contributed about 19% to the well-being of forest-based households during the COVID-19 pandemic.

Discussion
The empirical data obtained and analysed disclosed that many households responded to the socioeconomic impact of the COVID-19 pandemic by diversifying into forest-based activities such as fuelwood harvesting, extraction of medicinal plants, fruit-gathering, and beekeeping while others intensified these activities. This is precipitated by the demand for nutrition, income and health during the pandemic. This indicates that many households in and around forested communities will continue to rely on forest resource extraction as a safety net presently and also in the future given the lack of alternative non-forest based livelihood strategies in such communities. Significant diversification into medicinal plants is logical, considering the propagation of information during this period that certain medicinal plants have preventive and curative properties. Furthermore, most households depend on fuelwood for energy. Therefore, with more in-migrants received from the city during this period, the demand for fuelwood, fruits and honey logically increased.
Across 30 villages in the Stung Treng province of Cambodia, Nguyen et al (2020) also reported that health shocks intensify poverty among households. Hence, many households diversified while others intensified the extraction of forest resources such as fuelwood, game, vegetables and fruits. Diversification into fuelwood and medicinal plants extraction and beekeeping correlate significantly with household well-being while only medicinal plants intensification contributed significantly to household well-being. Fuelwood demand increased during this period as many people spent time at home and used more household energy (fuelwood). Also, as some households diversified into beekeeping as an alternative source of income to abate economic hardships. The case of medicinal plants is logical, as many households made use of these plants with the intention of benefiting from their preventive and/or curative properties towards COVID-19, despite the lack of clinical evidence to validate their conviction. In the context of Ethiopia and six villages in Mozambique and Malawi, fuelwood and other forest-based products supported household well-being by contributing to more than 30% of household income (Barany et al 2005, Melaku et al 2014. However, diversification into forest resource extraction creates negative externalities which adversely affect the wellbeing of forest-based households in other contexts (Razafindratsima et al 2021).
The logistic regression (see table 3) showed that the socioeconomic characteristics of forest-based households significantly influence the likelihood of livelihood diversification and intensification. Specifically, male-headed households are less likely to diversify into fuelwood harvesting, medicinal plant extraction and fruit-gathering during the COVID-19 pandemic. This could be attributed to the fact that some activities including fuelwood and medicinal plant harvesting are largely performed by females (Mosa 2016). In Northern Ghana, females were more likely to diversify into resource extraction during shocks than males (Zakaria et al 2019). However, the results contradict the findings of Demissie and Legesse (2013) in Ethiopia where males diversify into forest-based activities more than females. Furthermore, educated households are less likely to diversify into forest-based activities including the extraction of non-timber forest products. This indicates that education also enables households to access alternative livelihood strategies during shocks (Teshome and Edriss 2013). Large-sized households with many of their members being dependents, diversified into fuelwood harvesting, medicinal plants extraction and fruit-gathering. This is not surprising since such households have labour to support these activities. Equally, large household sizes signal an increased burden, this compelling household heads to find ways of catering for their young ones (Adepoju and Obayelu 2013).
However, there was an inverse relationship between access to credit and the preference for diversification into forest resource extraction during the COVID-19 pandemic. This follows the argument that financial credit gives households a wide range of non-forest-based livelihood opportunities to explore (Yunus 2007). Hence, a logical way forward is to provide microcredit in forest-based communities and to also make access flexible and at a low cost (interest on credit) for households. This can help to reduce poverty, increase resilience and/or divert livelihood attention from forest resources during current and future shocks. The study further reports that migrants were more likely to diversify into or intensify fuelwood and medicinal plant extraction during the COVID-19 pandemic than natives of the study area. Migrants in this area are financially constrained households originating from the Northern region of Ghana. They work as on-farm labourers for the natives. The onset of COVID-19 relatively intensified their financial handicap and hence, forest resources served as a salvage portfolio. This corroborates with Logie et al (2021) who found in the Bidi Bidi refugee camp in Uganda that migrants mostly engage in fuelwood and medicinal plants extraction for both economic and health reasons during the COVID-19 pandemic.
On the asset determinants of shifts in livelihood strategies, financial assets (access to credit), physical assets (ownership of mobile phone) and social asset (organizational membership) significantly predicted the likelihood of households diversifying into forestbased activities during the pandemic. Also, monetary savings (financial assets) and land ownership predicted the likelihood of forest-based livelihood intensification. This implies that diversification or intensification choices during shocks including the current pandemic hinge around household assets as theoretically argued (see figure 1). Precisely, households may have accessed the market opportunities of some forest-based products including medicinal plants and fruits by connecting with their buyers through mobile phones. This explains why households with mobile phones diversified in forest-based livelihood activities more than their counterparts during the COVID-19 pandemic. Studies across Northern Ghana and Asia reported similar findings whereby mobile phones played a critical role in the food supply chain as well as in the marketing of forest products during the COVID-19 pandemic (Zakaria et al 2019, Belton et al 2021. Also, the high forest-based livelihood diversification or intensification among households with monetary savings of less than 500 dollars reiterates the position that the poor are likely to engage in resource extraction more than the better-off households during shocks. Similar studies in Eastern Honduras announced that wealthier households with high savings are less likely to engage in forest-based activities as compared with poor households with fewer savings (McSweeney 2004). On the contrary, it is reported that betteroff households are more likely to engage in forestbased activities including timber sales than the poor since this requires a high amount of investment (Mamo et al 2007). We however argued that these high return forest-based activities demand a relatively high amount of investment. Hence, the poor diversify more into forest-based activities, not for capital accumulation but for their day-to-day survival as most households asserted that they engage in these activities for subsistence. This means that strategies to elevate poorer households around forested communities in the Bia West District of Ghana and other areas in SSA, as well as the global south, could enhance forest conservation. Studies that uncover the conservation implications of livelihood dynamics should be considered in future.

Conclusion and policy recommendations
Diversification into forest-based activities was more prevalent among forest-based communities in the Bia West District of Ghana during the pandemic than intensification. The activities that households diversified into include fuelwood harvesting, extraction of medicinal plants, fruit-gathering and beekeeping although some others rather intensified. While increasing diversification could enhance livelihood resilience during shocks, it may have adverse implications in terms of the reduction in the availability/ abundance of forest resources. Sustainability could be compromised in this regard, especially for medicinal plants, fruits and fuelwood. However, diversification and intensification in beekeepinga conservation-friendly livelihood strategycontribute to the sustenance of forest resources. Shifts in livelihood strategies hinge on the assets and socioeconomic characteristics of households including gender, household status (migration), education and income status. As this paper did not report on the differentiated motives behind diversification (e.g. domestic or commercial), future studies should consider investigating such motives in the context of pandemic-related shocks. In this study, shifts in forest-based activities were categorised under diversification, intensification and business as usual. Consequently, a potential fourth option-diversify and intensify-was not considered. This constitutes a limitation to the study. Future studies should therefore incorporate other categories including households that may have diversified and intensified.
Among the forest-based livelihood activities that households diversified in, fuelwood and medicinal plants extraction significantly correlate with the wellbeing of forest-based households. Also, medicinal plants intensification correlated directly with household well-being. That is, these activities lead to an improvement in well-being captured to include income, savings, health and the nutritional needs of forest-based households during the COVID-19 pandemic. As this paper reported aggregates of wellbeing (changes in income, food availability, health, and savings), further studies are required to establish the specific well-being improvements linked to either diversification or intensification. The diversification or intensification of forest-based activities could compromise the conservation goals around this area. While we call for future studies to establish the conservation implications of diversification and intensification, we submit at the moment, that livelihood activities should be significantly regulated to ensure that poor and marginalised households survive while also assuring conservation during shocks such as COVID-19. Additionally, the government and agencies of conservation interest should embark on livelihood resilience and capacity building programs around forested communities as a mechanism to improve household well-being during the COVID-19 pandemic and act against possible shocks in the future.
With this submission, we advocate that in the long run, policies to incentivise forest-based households to access training for the domestication of certain forest resources (e.g. fruits and medicinal plants) should be encouraged to serve as a buffer in times of increased usage such as during pandemics. This could be assured through the provision of credit facilities to target communities. This is a logical way forward to provide alternative livelihoods to forestbased households since the continuous reliance on forest resources can lead to degradation if not technically managed. This paper only focused on analysing the dynamics introduced in the forest-based livelihood of households in forest communities during the COVID-19 pandemic. Hence, future studies should consider incorporating non-forest-based livelihoods. Also, the investment required for the key forestbased livelihood strategies that forest-based households diversified into or intensified can provide an entry point for future studies.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).