The study was conducted at Belete forest southwest Ethiopia. Geographically, it is found between 36° 15′ E and 36° 45′E and 7°30′ N and 7°45′N (Figure 1). Belete forest belongs to the moist evergreen Afromontane forest of southwest Ethiopia. The forest is one of a few remnant Afromontane moist evergreen forests in southwest Ethiopia. Belete forest, together with Gera forest, was designated as one of 58 national forest priority areas in Ethiopia in 1989 (Cheng et al., 1998). The study area is characterized by a mosaic of forest, cultivated land and settlements. The most accessible area is managed for coffee production involving planting of wild coffee taken from coffee forest and intensive (under growth removal and canopy reduction) management for coffee agroforest. The forest has been under participatory forest management for the last two decades. The forest is divided into blocks of forest among the forest user groups. Forest is a source of livelihoods for people living in and adjacent to the forest. The present study worked with five forest user groups namely: Dabbiyee, Gurrattii, Qartammee, Mexxii-Caffee, and Sokii forest user groups. The total number of households within a village are in the hundreds. The dominant ethnic group are the Oromo, most of whom are Muslim with a few Christians. The local people organized into forest user groups and signed an agreement with Oromia Forest and Wildlife Enterprises to be entitled in accessing and using forest resources. Forest resource use pattern of the local people changes with time. Currently, the tradition of forest resources use is dominated by coffee production.

Both ecological and ethnoecological data were collected (Lucena et al., 2013; Lokonon et al., 2019; Ribeiro et al., 2019; Korach et al., 2020). The data were collected in two steps. First coffee agroforest inventory was carried out to collect ecological data in five sites (Dabbiyee, Gurrattii, Qartammee, Mexxii-Caffee, and Sokii) and then coffee agroforest owners were interviewed on the use and preference of woody plant species in coffee agroforest. A total of 63 plots (400 m²) (Dabbiyee = 15, Gurrattii = 12, Qartammee = 11, Mexxii-Caffee = 12, Sokii = 13) were selected for woody plant species inventory. It covered a total area of 2.52 ha. The plots were laid systematically along the transect in coffee agroforest in each site. Within 20 m × 20 m, all woody plant species with diameter at breast height ≥ 10 cm were identified and recorded. The researcher identified woody plant species in the field by their local names with the help of local people and cross-checked using available literature (Bekele-Tesemma, 2007; Eyasu et al., 2020). The specimen of woody plants difficult to identify in the field were collected and taken to lab for further identification with the help of botanist. Plant specimens were deposited at Jimma University Department of Biology. For regeneration assessment, seedlings and saplings were identified, counted and recorded within a sub-plot of 10 m × 10 m and 5 m × 5 m, respectively. All methods were performed in accordance with the relevant guidelines and regulations. Plant identification was done following the flora of Ethiopia and Eritrea.

Ethnoecological data were collected through semi-structured interview. A checklist for an interview was prepared focusing on the use and preference of woody plant species in coffee agroforest. All woody plant species recorded in the field were included in the semi-structured questionnaire. Free listing technique was used to record the use of each woody plants. Coffee owners list as much as they can remember the use of the plant (Martin, 1995). The use of woody plant mentioned by interviewees were grouped into different use categories. Moreover, coffee owners were asked to mention the most preferred coffee shade trees and their management practices. The ethnoecological data were collected from 96 individuals (11 females) (Dabbiyee = 20, Gurrattii = 18, Qartammee = 17, Mexxii-Caffee = 21, Sokii = 20). The age of interviewees ranges from 20 to 80. Permissions were obtained from Oromia Forest and Wildlife Enterprises, Shabe Sombo district office and the lowest administrative Office (kebele) to undertake the study. All methods were performed in accordance with the relevant guidelines and regulations. The interviewees gave their consent on verbal than written form for an interview. To minimize the bias due to peer interference, an interview was carried out on an individual based on the convenient time and place to interviewee.

Alpha diversity and other indices were computed for diversity assessment using PAST version 4.03 software. Alpha diversity is expressed as the total number species (species richness) in coffee agroforest (Manaye et al., 2021; Marzialetti et al., 2021).

Species richness was computed using the formula:

where ni is the number of species in a coffee agroforest.

Woody plant species preference in coffee agroforest was analyzed using the number of citation given to each woody plant species for the respective use categories. Citation refers the number of use of wood plants the interviewee mentioned (Lucena et al., 2013; Lokonon et al., 2019). Literature has stated that more preferred woody plant species are more cited (Lokonon et al., 2019). Woody plant preference in coffee agroforest was estimated using the equation adopted from Duguma and Hager (2010) as follow:

Where M S c s p p x , use ( y ) stands for mean citation score of species x for use type y, n stands for the total number of interviewees (n = 96); ARPS s p p x , use y stands for the adjusted relative preference score of species x for use type y in % and AGRPS s p p x , use y stands for the aggregate relative preference score of a species across all types in percentage. Aggregate relative preference score was computed for multiple use and shade value of recorded woody plant species in coffee agroforest.

The woody plant conservation priority (CP) analysis adopted with some modification the technique that was employed by scholars (Dzerefos and Witkowski, 2001; Oliveira et al., 2007; de Albuquerque et al., 2011; Lokonon et al., 2017; Kafoutchoni et al., 2018; Ribeiro et al., 2019; Oliveira et al., 2021). Table 1 portrays the criteria and score employed in the analysis. Woody plant species is calculated using the formula:

Where CP corresponding to Conservation Priority, BS corresponding to biological score estimated based on relative density (D) as BS = D × 10. The usage risk (UR) is estimated based on management risk and use value (U) as UR = [0.5(H) + 0.5(U)] × 10. Use value is estimated as the average of the sum of the local importance (L) and the diversity of use (V) (Ribeiro et al., 2019). For woody plants that have timber and construction value additional value of 10 points were added as additional usage pressure (Ribeiro et al., 2019). Finally woody plants divided into three categories, category 1 with high priority species for conservation (CP ≥ 85), category 2 with moderate priority species for conservation (60 ≤ CP < 85) and category 3 with low priority for conservation (CP < 60). Spearman correlation was computed to test the relationship between woody plant species preference and conservation priority.

Findings on ecological data showed that many woody plant species associated with coffee agroforest in southwest Ethiopia. The result showed that 48 different woody plant species belonging to 27 families were recorded in 63 plots (Table 2). Most of the families were represented by a single species. Only a few family consists of a maximum of four species. Among the recorded woody plants three woody plant species, Cordia africana, Croton macrostachyus, Milletia ferruginea were more abundant compared to the other species accounting for 41 percent of the number of woody plants. Forty five woody plants had contributed each less than 5 percent of the total abundance. The lower abundance of many woody plants were the outcome of coffee management that resulted in stem reduction. Muleta et al. (2011) have reported the family Fabaceae dominate coffee agroforest in southwest Ethiopia. Aerts et al. (2011) reported Croton macrostachyus and Milletia ferruginea dominate coffee agroforest in southwest Ethiopia. This is attributed to the regeneration characteristics of individuals (Aerts et al., 2011). A Study from Dallo Mena district, southeast Ethiopia has reported 10 different tree species in shade grown coffee (Mengistu and Asfaw, 2016). Another study from Jimma area southwest Ethiopia have reported 38 different tree species in coffee agroforest (Worku et al., 2015).

Higher species diversity with Fisher alpha 12 and Shannon Weiner diversity (H) 3.08 were found in coffee agroforest (Table 3). Previous studies categorized Shannon Weiner diversity as high with a value ≥3, medium with a value between 2 and 3, low with a value between 1 and 2, very low with a value <1 (Atsbha et al., 2019; Fentaw et al., 2022). A Shannon Weiner diversity value of 3.08 of the present study belongs to a high diversity category. The individual based rarefaction curve showed the estimated number of species as more number of individuals recorded (Figure 2). The Choa-1 value of 51.5 showed the maximum species richness estimated with more sampling effort (Table 3). Worku et al. (2015) have reported Fisher alpha diversity of 8.53 in coffee agroforest in Yayu southwest Ethiopia. Kewessa et al. (2019) have found a Shannon diversity of 1.74 in coffee agroforest Bale Eco-Region, southeastern Ethiopia. Senbeta and Denich (2006) have reported a Shannon diversity of 2.82 at Bebeka southwest Ethiopia. Rigal et al. (2018) has reported a Shannon diversity of 3.42 with 30.57 effective number species in coffee agroforest from southwest China.

Coffee agroforest is a source of livelihoods for the local people. It provided ecosystem services that benefit the forest users (Bukomeko et al., 2019). The present study showed that woody plants maintained in coffee agroforest provide diversity of uses. Ten uses such as fuelwood, charcoal, Construction, medicinal, coffee shade, bee forage, beehive, farm tool, hanging beehive and timber that determine the management of woody plants species in coffee agroforest were frequently mentioned. These uses can be destructive (timber, beehive, construction, charcoal, farm tool), partial destruction (fuelwood, medicinal) and non-destructive (coffee shade, bee forage, hanging beehive). Ecological and economic reasons are the driving factors for woody plant management in coffee agroforest. In coffee agroforest the shade value of woody plant species are the priority for tree selection and management. Nevertheless, the present study findings showed that coffee agroforest owners obtain multiple benefits from the managed woody plants. Girma et al. (2019) have stated that local people manage woody plants for construction, fuelwood and honey production. A study from Bangladesh showed that local people manage woody plants for multiple uses and the major uses are fruit, fuelwood, pole, timber, medicinal etc. (Tarit et al., 2015).

The aggregate relative preference score (ARPS) showed 15 most preferred woody plant species in coffee agroforest (Table 4). Each woody plant species provided multiple uses and the relative importance differ between the species. Based on the all uses, P. adolfi-friederici, C. africana, P. fulva, E. candelabrum were the most preferred woody plants. The use value of these woody plants were mentioned more for timber, hanging beehive, beehive than coffee shade. A. gummifera and M. ferruginea were the most preferred coffee shade trees. The abundance of P. adolfi-friederici, C. africana, P. fulva, E. candelabrum were lower than M. ferruginea a well-known coffee shade tree in southwest Ethiopia. A study from Tanzania has shown local people give priority for the tree species that provide food, fodder and fuelwood (Wagner et al., 2019). Bukomeko et al. (2019) have studied the relationship between tree diversity and farmers need for the benefit of trees and found that farmers need did not match with tree diversity in coffee agroforest in Uganda. Lamond et al. (2016) have investigated underpinning factors for tree preference in coffee agroforest and reported that multiple uses (both ecological and socioeconomic) determine the tree selection in coffee agroforest. Albertin and Nair (2004) a have studied farmers’ perspective on the role of shade tree in coffee production systems in Nicoya Peninsula, Costa Rica and have found tree species that are not preferred for coffee shade still maintained in coffee agroforest for the benefits they provided for the local people. The same author highlighted the need for incorporating more trees and fruit trees in coffee agroforest in the region. A study by Hundera (2016) has shown local people maintain Schefflera abyssinica and Olea welwitschii in coffee agroforest for honey production in southwest Ethiopia. Valencia et al. (2015) have reported bulk of tree species that have not been valued for coffee shade in coffee in Chiapas, Mexico. Reinforcing reasons that encourage tree management in coffee agroforest are the need for additional benefits such as timber, fuelwood, medicinal and other non-timber forest products (Valencia et al., 2015). A study by Kalanzi and Nansereko (2014) has shown local people in Bukomansimbi district of Uganda prefer tree species that provide multiple products in coffee agroforest.

The present study findings showed that the conservation priority (CP) varies between woody plants. Three types of categories were identified for local conservation priority that took into account the management practices, utilization and regeneration status of each woody plant (Table 5). Category 1 indicates woody plant species that need high conservation priority and accordingly category 2 and category 3, moderate and low priority, respectively. As indicated in Table 5, 12 species are represented under category 1, 19 species are represented under category 2 and 17 species are represented under category 3. Woody plants that are destructively utilized and removed from coffee agroforest during slashing under growth plants belongs to category 1. These woody plants had a few individuals and insufficient regeneration. Local people do not value these woody plants for coffee shade and totally remove, if possible, from the system. Likewise woody plant species under category 2 are utilized destructively that resulted in low number of individuals leading to loss of the plants in the long run. The Spearman correlation showed negative correlation between abundance and conservation priority [r_s(46) = −0.681, p = 000]. Most preferred woody plants such as Cordia africana, Polyscia fulva, Pouteria adolfi-friederic and Olea welwitschi belongs to category 3, Euphorbia candelabrum, Ekebergia capensis, and Fagaropsis angolensis belongs to category 2 and known coffee shade tree species Milletia ferruginea and Albizia gummifera belongs to category 3. This work is the first attempt to classify woody plant species in coffee agroforest in southwest Ethiopia. It highlights the status of woody plants under coffee management practices. The study findings complement the notion coffee agroforest is tree diversity conservation hotspot (Valencia et al., 2014). Local farmers knowledge plays a decisive role in conservation of tree species in coffee agroforest (Valencia et al., 2015). Joshi et al. (2019) has stated that woody plant species recognized as useful are under pressure for utilization and need attention for conservation. In the present study Cordia africa which is extracted for timber is exceptional due to the nature of plant regeneration characteristics. Cordia africana is found in low abundance but withstand timber utilization. Joshi et al. (2019) has reported tree species require high conservation priority compared to shrubs. But, the present study showed that shrubs require more attention than trees as coffee management remove under growth including shrubs through slashing. The study also support Lokonon et al. (2017) that state most used species are not top priority for conservation. For instance, in this study Cordia africana is highly utilized for timber but categorized under category 3. Rytigynia neglecta, Maytenus arbutifolia and Ilex mitis are among the species with high diversity of uses but totally discouraged in coffee agroforest in southwest Ethiopia.