Effects of Coffee Management on Bee Floral Diversity, Honey Yield and Quality: The Case of Gera District, Jimma Zone, South West Ethiopia

This study was conducted to assess the effect of forest management for coffee cultivation on bee flora diversity, honey yield and quality as forest crop products in Gera district. Forest coffee (FC) and Semi-forest coffee (SFC) plots were selected for bee flora assessment. Total of 34 plots (FC =17, SFC =17) with plot size 20m X 20m (400m 2 ) were assessed. Sixty bee flora species belonging to trees (30), shrubs (21) and woody lianas (9) were identified and compared across plots. Results show that more bee flora diversity in FC (2.03) than SFC (1.09) system. Honey yield data was collected from 78(FC=52, SFC=26) beekeepers. The honey yield of FC was higher than SFC system in both high and low production years. The honey production on average was 9.58 Kg/hive for FC and 6.44Kg/hive for SFC in high production year while 6.5 kg/hive for FC and 4.24 Kg/hive for SFC in low production year. To assess the honey quality 6kg honey samples (FC=3, SFC=3) were collected. Six honey quality parameters i.e. moisture, ash, pH, free acidity, electrical conductivity (EC) and total soluble solid (TSS) contents. The biochemical variation in the composition is significantly different (P<0.05) in ash content, pH, EC and free acidity when comparing FC with SFC honey samples while the percentage of moisture and TSS contents were insignificant (p>0.05). The study revealed that; FC conversion to SFC is associated with a decline in bee flora diversity, honey yields and on top of this, it has implications on honey quality deteriorations. Thus, there is an urgent need for control and monitoring on the expansion of SFC cultivation, which needs immediate conservation measures. Therefore, conservationists have to find other means of economic incentives for the community to refrain from further FC conversion and intensifications.


Introduction
Ethiopia has huge potentials for beekeeping given the prevailing suitable ecological conditions and floral diversity, thus making it highly suitable for sustaining large numbers of bee colonies and the long established practices of beekeeping. As of 2015, the country is the top producer of honey in Africa while ranking amongst the top ten in the world. It accounts for 23.6% of the continent's total honey production (McGill 2016).Forests have a potentials to provide adequate bee-forage in terms of both quality and quantity of nectar and pollen grains. For this reason, beekeeping has also the potential to increase opportunities for forest conservation. When promoted among forest adjacent communities, beekeeping provides reliable livelihood options (Deffar 1998).
In spite of the suitable ecological conditions and floral diversity, Ethiopia produces about 163,257.42 metric tons of honey in -2011(McGill 2016,but it has the potential to produce 500,000 tons of honey per year (MoARD 2008;Ayalew 2008). The low honey production in Ethiopia can be attributed to the dominantly smallscale operations, which employ traditional methods of production. Moreover the human encroachments and modification of natural forests, the bees and the plants they depend on, are constantly under threat (Bahru 2017).
Forest coffee (FC) and honey from specific forest types fetch high premium which are plant crop products. It is several of the widely cultivated forest-derived cash crops (Wiersum et al. 2007). It has acquired significant economic importance in the country as commonly harvested from natural forest (FAO 2016) particularly in south westernt(SW) part of Ethiopa.Coffee is one of the most imporatnt cash crop plants that farmers need to miximize its production through traditional forest management practices ( Gole 2003).According Hwang, Hundera, and Mekuria (2020) findings suggests that the expansion of areas of intensive management of the coffee forest and the intensification of the consequent degradation of that forest occur in the study area due to rising of coffee prices. These leads to the gradual modifications of FC in to semi-forest coffee (SFC) systems in the course of removal of trees and under growth vegetations. The results from the tie-point method suggests that the expansion of areas of intensive management of the Thus, frequent clearing of small tree, shrubs and climbers in managed forest has negative effects on structure and composition of communities and species diversity in the forest including regeneration of tree species and coffee itself (; Senbeta and Denich 2006;Gole 2003). These indicates that there is trade-offs between maximizing production and maintaining the forest and its biodiversity (MEFCC 2018).
Furthermore as coffee management continues, the FC could be changed to coffee farms with a few shade trees (Kufa 2010) resulted in a loss of forest-based woody species including important bee plants once FC are converted into SFC systems (Tadesse et al. 2014).Thus,the change from FC causes a reduction in bee floral diversity,ultimately affecting the quantity and quality of honey as forest crop productions. Therefore, a comprehensive study is needed on the effects of coffee management on bee floral diversity, honey yield and quality of Gera district, as well as SW part of Ethiopia. This was an area of investigations that has been neglected and yet holds significant potential for future sustainable FC management initiatives and maintains forest derived crop products.

Descriptions of study area
The study was conducted in Gera District Jimma Zone of Oromiya Region, South-west Ethiopia. It is located within the longitudinal range 35° 57′ and 37º 37′ East and latitudinal range of 7° 13′ and 8º 56′ North (Figure 1). The mean annual maximum and minimum temperatures are 24.2 o C and 14.2 o C, respectively, and the mean annual rainfall is between 1,880 and 2,080 mm. The major soil types are: Arcisol, Nitisol and Leptosol (Tulu et al. 2014).The remnant forest vegetation at Gera area can be categorized as tropical Afromontane moist forests which have been further classified into: natural forest (virgin and disturbed) and plantation forests (Tadesse et al. 2014). Within this moist, shaded curtain of vegetation, Ethiopia's rich varieties of coffea arabica evolved in a wild. It is one of the remnants of broad leaf moist forest in Ethiopia comprising economically and ecologically important plants (Mohammed and Bekele 2014).The vegetation cover of the area was estimated to be 56% of the total area.

Study site selection
For this study, FC and SFC forest with traditional bee keeping activities were considered based on the intensity of forest management practices. Here in FC system, coffee is harvested directly from spontaneously regenerating natural population of coffee. The only management practice in the forest system is access clearing to allow movement in the forest during harvesting time while SFC represents a system in which the forest is managed or manipulated mainly for coffee productions. In this system, small trees and shrubs competing with coffee are cleared. Clearing is twice a year, one before harvesting season and another after harvesting, before the main rainy season starts (Tadesse 2003). The difference here is the intensity of management practices. The selected SFC was about ten years old since they were converted from FC system and FC conversion under smallholder farmers were considered for this study. 2.2.2. Bee flora assessments Prior to bee flora sampling, a systematic sampling design was used to collect data on bee flora and environmental parameters. In each system, quadrants of (20m X 20 m) (Senbeta and Denich,2006) was laid down along transects at 500 m distance apart with each plot 300m far apart from one another. All transect walks were taken from south to north across each forest block then after the second plot transact walk turned to West direction uniformly for all plots. A total of 34 quadrants (Dallmieir 1992) was laid; FC ( n = 17) and SFC( n = 17 ). In each quadrant, all bee flora species belonging to (trees, shrubs and woody lianas/climbers) growing habits were counted. Because those are most affected species by management for coffee production, even though there is most important herbaceous bee plant are available. Then scientific and local name of identified bee flora species were triangulated (Bekele 2007;Edwards 1995;Reinhard 1994) The field survey works were focus to assess common bee flora species abundance, diversity (Shannon diversity indices ((H`)), species richness and species evenness (E) were computed. Because of these techniques analysis can help elucidate qualitative differences among the sites from multi vitiate data sets. These common bee floras were once identified during the social survey (Focus group discussions)(( FGD)) in local language (Afan Oromo) were recorded during the plot survey with the help of knowledgeable beekeepers participated on FGD, and Honey bee flora of Ethiopia were used for identifications of species difficulties after photo captured during field work (Reinhard 1994;Cfdr 2010 Household(HH) Survey Focus Group Discussions (FGD): were conducted in the study area with purposively selected section of community such as 'Walda Bulchinsa Bosona' (WaBuB) ((Forest Management Association)) members, development agents (DAs), beekeepers, and some individuals, who were believed to be knowledgeable about bee flora plants in the forests. Hence, purposive sampling method was used for selecting FGD members. This work was used mainly for identifying common bee flora of study area and furthermore to pinpoint the situations of FC conversion activities in relation to their honey productions.
To select the number of samples of HH for the study first number of HH who practice beekeeping alongside with FC management of each selected kebele were listed from WaBuB community. In both sites, there were seasonal and residential forest users here after (beekeepers) that have their owned individual forestland known as 'mume' hereafter (forestland used for traditional beehives installations). However, the collected data was from residential users only. After getting the total number of HH fills the criteria in each selected localities, the sample size was determined using probability proportional to sample size-sampling technique (Bartlett 2001).
Probability proportional to size approach was used to determine the number of sample HH from identified beekeepers for each study site. Single HH respondent was used as sampling unit in this study. Accordingly, the total numbers of HH head fill the criteria living in both localities were 206 (135 from FC and 71 from SFC). From all the identified, the HH head (beekeepers) having either FC or SFC forest type was systematically taken for a formal interview. Based on Cochran (Bartlett ,2001) population correction factors, a total of 82 sample HH head were selected from the total bee keepers. Allocations of the number of sample HHs to each study site, was proportional to the number of HH head participated on beekeeping in each selected site, accordingly, 54 HH from FC and 28 HH from SFC were selected for this study. Only the first respondent was chosen randomly. Hence, to collect information regarding bee forage plants and related parameter (like identification of common bee flora, bee keeping in relation to FC and SFC management, consequences of FC conversion on honeybee floras and honey yields. etc.) for the study, the sampled beekeepers were individually interviewed with structured questionnaire. The coffee management practice questions included shade tree selection criteria and level of forest conversion, intercropping accessibility of bee floral sources, preferences of where they are deploying their beehives, numbers of traditional beehives of each and application of pesticides in SFC(if any)was identified. 2.2.4. Key informant interviews: Key informant interview was made with Gera district beekeeping expert and agricultural office expert specially for selecting the potential kebeles for the study, DAs of the study area, WaBuB community members, forest expert and some individual beekeeper and knowledgeable person who can be used as guide person during plot inventory in both forest types. The qualitative information collected in interview was used to supplement and crosschecked the data obtained through the HH survey. Hence, purposive sampling method was used for selecting members for the key informant interviews . 2.3. Data analysis 2.3.1. Bee flora species composition and diversity analysis The abundance of bee flora species defined here as the total numbers of all individuals' bee plant species in all 34 quadrants were estimated for each management system (FC and SFC). The relative frequency of each bee flora species that are top ranked among most important bee flora not among the all surveyed species (according information obtained from FGD)were calculated.
The Shannon diversity indices for the common bee flora species in the study area were estimated in the two different forest management systems (FC and SFC). To compare bee flora species composition of the two forest types; species richness, Shannon diversity index, and Shannon evenness index were calculated by common biodiversity indices formulations (Senbeta and Denich 2006;Dallmieir 1992).
Furthermore, the similarity of bee flora species between habitats (FC and SFC) was also calculated by Jaccard similarity coefficient (Jaccard 1912).
To summarize the overall obtained result i.e. bee flora species abundance, frequency, and diversity, the statistical computations were made by Microsoft (excel word 2010) and the result were presented in the form of tables.
Furthermore the distributions of bee flora species between the plots of the two forest types (FC and SFC) were evaluated by using χ 2 -tests (chi-square) with SPSS version 23 to test whether the FC conversion bring significant effects on bee flora compositions between expected and observed bee flora species recorded. Then the data's were interpreted. 2.3.2. Honey production data analysis It was analyzed as an average production in (kg per hive) for both forest types in both high and low production years that obtained from respondents. The data collected were summarized by using descriptive statistical methods (such as frequencies, percentage and graphs) and the obtained results were presented in the form of tables and figures. The statistical computations were made by Microsoft (excel word 2010). 2.3.3. Honey quality analysis The honey quality, which expressed as physicochemical properties of the honey, was determined. During primary data collection a total of 6Kg of honey samples, (n=3Kg) from FC and (n=3kg) from SFC were considered taking an account as bees can be forage a radius of 6000 m and an area of 113 km 2 away from their hives (Couvillon et al.2016).Thus, the two sites were more than 25 km far apart and the forest coverage's were 3774 ha, 811ha for FC, and SFC respectively (OC. WaBuB.2008).Hence, there is no bias of mixing up of the taken honey samples by bees. The 3 x 3 factorial arrangement replicated thrice was used. All collected honey samples were fresh that directly taken from bee keepers(traditional hives) during the peak honey harvesting season(March to April) and coded as honey samples from FC site,(FC-HS1,FC-HS2,FC-SH3) from SFC site,(SFC-HS1,SFC-HS2,SFC-HS3) and packed with plastic bottles with their specifications of harvesting days. There were no differences in harvesting seasons of all collected honeys samples. Hence the physicochemical compositions (moisture content, ash contents, pH, free acidity, EC, TSS) of honey samples were determined according to the Harmonized Methods of the International Honey Commission (HMIHC) and Ethiopian honey quality Standard (Brno 2007;ES 2005). All physicochemical properties of the collected honey samples except moisture content were analyzed at JUCAVM, Department of food Science and Post-harvest Management and NRM department at Soil Science laboratories. Because of those laboratories were works on (ash contents, pH, free acidity, EC, TSS) parameters only the work was limited to these parameters. But the moisture contents of honey samples were estimated by modern Automatic Temperature Compensation (ATC) versatile refractometer model number M106401 at farm gate. The results were subjected to analysis of variance (ANOVA). All statistical computations were made by using SPSS version 23-computer software in order to determine whether there were significant differences in quality existed for selected parameters between honeys harvested from both sites. Differences between means at the 95 % (p≤0.05) confidence level were considered as significant differences.

Demographic characteristics of respondents
From the total of 82 samples HH selected, 78 respondents (92) % were interviewed. The average age of the respondents were 33.68 years that about 69.4% of them lies between 25 to 40 years with average bee keeping experience of 13.84% years and about 80.2% of them acquired habits of beekeeping from their parents while the rest 19.8% of them acquired it through motivation (interest) of themselves (see Table [1]). The result showed that different age groups can perform beekeeping and in most cases, people at younger and older ages are more engaged in forest beekeeping. The short summaries of demographic characteristics of respondents are shown as illustrated in Table [

The common bee flora species in FC and SFC of Gera district
A total of 60 bee flora species belonging to 35 families were recorded for both (FC and SFC) systems (see Table  [2]). This shows that Gera forest contains a high number of bee flora species that are remarkable for honey productions. The recorded bee flora species in present study were higher than the previously reported ones by (Senbeta et al. 2013) which revealed that 32 plant species representing 19 families were recorded as being sources of honeybee forage in coffee forest of Ethiopia ; Yayu, Bonga,Harenna and Sheko forest. The most common and important bee flora species in the study area according to FGD conducted includes Schefflera abyssinica,Croton macrostachyus, Olea welwitschii, Pouteria adolfi-friederici , Syzygium guineense,Teclea nobilis,Vernonia amygdalina,Vernonia auriculifera. This is supported by (Gezahegn et al.2014) who conducted the same study area states that this bee forage flowers at different times of the year and thus offers a possibility of harvesting honey up to four times per annum in some cases. This agree with (Ito 2014 which revealed that beekeepers place their traditional beehives where those tree species are dominated due to their rich nectar and can produce quality honey. This is also in line with (Amado 2015) which states those plant species are the dominated tree and shrubs in Gera forest. Therefore, it indicates that the different bee flora species in the study area were contained in different family species. The growth habits of recorded bee flora species in the study area, were characterized as tree for example (Schefflera abyssinica,Syzygium guineense),shrubs (Vernonia amygdalina,Vernonia auriculifera),wood liana (Clematis simensis, Embelia schimperi,Hippocratas africana)(see Table [ 2]).From all the species recorded 30 (50%) were trees, 21(35%) were shrubs, 9(15%) were woody lianas (see Table [2]).Hence, trees were more dominantbee forage followed by shrubs and woody lianas respectively in both forest types. Forest resources in SW are mostly harboring diversified tree species. This may be the main reason for the dominance of bee flora species belonging to tree growth habits in the study area. This is also further supported by Gebrehiwot and Hundera (2014) that states the growth forms of the species recorded from Belete moist evergreen montane forest was dominated by herbs, followed by trees.  et al. 1993;Fichtl and Admassu Addi, 1994;Edward 1989;www.theplanlist.org) 3.2.1. Bee flora species abundance and diversity in both forest management systems (FC and SFC)

Bee flora species abundance
The survey result showed that, total number of bee flora species was higher in FC (57) system than SFC (50) system (see Table [ 2]).It indicates that the coffee forest has high flora resources. According to ( Gole et al. 2015;Gole et a.l 2008.) that states coffee forests have, to be viewed as a complex mosaic of different plant communities. However, in both forest management systems there were unique bee flora species recorded. In FC system about 16.6% floral species were identified and about 5% of them were exclusively found in SFC sample plots. Nevertheless, except Fagaropsis angolensis and Dombeya torrida the rest species were observed outside of the plots of SFC site indicated that could be grown in forest type, while only Senna septentrionali bee forage species was not observed in FC system. From the result the dominant bee flora tree species for FC includes; Syzium guineense, Olea welwitschii, and Schefflera abyssinica. The finding is in line with (Amado 2015) which states that those are dominant species in the tree layer of plant community types in Gera forest. According (Alemu 2012) the dominant tree species here in SFC includes; Albizia gummifera, Croton macrostachyus and millettia ferruginea. The finding is similar to Berecha et al. (2014) that states that those tree species are dominant in SFC site for their preferences of coffee shade. Regarding the abundance of bee forage most of individual bee plant species recorded in FC site was Coffea arabica, which was about 53.06%of the total species, while 80.34 %( see Appendix [1]) of Journal of Biology, Agriculture and Healthcare www.iiste.org ISSN 2224-3208 (Paper) ISSN 2225-093X (Online) Vol.10, No.16, 2020 was coffea arabica for that of SFC site. This indicated that almost all bee flora recorded in SFC site was coffea dominated. Thus, coffee domination indicates that the expansion of conversions of FC in to SFC system for coffee cultivation was very high.
Furthermore, there were bee flora tree species debarking (see Fig. [2]) observed in SFC management system during plot inventory. This the reason why important bee flora was less reduced in SFC system than FC systems. This showed that forest derived coffee corps became high prices in world market that is why farmers more intensified the forest to coffee framings by reducing shades. This showed that continued conversion factors for coffee cultivations linked with declining of bee forages abundance and its composition. Hence, the result clearly showed that FC land conversion affects the abundance of bee flora.
Relative frequency:-from the survey result complementary bee plants, the most ten leading species interims of their significances in honey productions due to their high floral nectars, were identified (see Table [3]). Even honey products are named after these species. Like Butoo from 'Schefflera abyssinica' (white honey),'Ibicha' from 'Vernonia amygdalina'(black honey),makkannisa from 'Croton macrostachyus ' and keraro honeys (light red) from'Pouteria adolfi-friederici'are recently most common honey types of the Gera district. Even if coffee was highly frequent in both forest types and important bee forage but did not included here because of its bloom not honey harvested commonly after this species. This is supported by Tulu et al. (2014) who conducted research at Gera states that those bee forage flowers at different times of the year and thus offers a possibility of harvesting honey up to four times per annum in some cases.

Bee flora species diversity in relation to both forest types
The computed Shannon diversity for bee flora in FC system (2.03) was higher than that of SFC system (1.09) (see Table [ 4]). Higher diversity in FC management system is an indication of the site had more plant species, since a Journal of Biology, Agriculture and Healthcare www.iiste.org ISSN 2224-3208 (Paper) ISSN 2225-093X (Online) Vol.10, No.16, 2020 greater variety of species allows for greater species interactions, while the continued FC conversion activities reduces the diversity of bee flora species resulting in coffee dominated in SFC systems. The finding is similar to Senbeta and Denich, (2006) which state that low Shannon diversity in the SFC systems is an indicative of the high abundance of one or a few species. Hence, the low diversity of the SFC system can be attributed to a large number of Coffea arabica individuals. According to (Gole 2003) (Gole 2003) (Gole 2003;Gole et al .2003) managed forests (SFC), diversity decreases with duration and intensity of management, the lowest being in the Semi-Forest-Plantations. These findings imply that the high intensive management in SFC for coffee production reduced bee flora diversity thus, conservation of forest tree species is a viable sustainability strategy from a biodiversity point of view, and that initiating smallholder beekeepers of FC management system is a viable economic activity to halt it from further intensifications. In this study species richness (S) was computed as, the observed number of bee flora species for each forest management system (see Table [4]). As a result, among identified 60 bee flora, the numbers of species observed in all plots of the FC were 57, which were relatively higher than those in that of SFC (50) system. The reason why FC was higher bee flora species richness and diversity than SFC is due to the less intensity of its managements. This is supported by Senbeta and Denich,( 2006) which revealed that, the dominance of species rich families rank also changes when the FC is converted into SFC system, reflecting the targeted removal of species. Thus, in this study 92.5 %( see Table [5]) respondents from FC management system confirms that there was only once per year clearing under coffee which is only during coffee collecting seasons. Meanwhile there was less human interventions and this creates good opportunity from natural regeneration of species as well as good vegetation cover; this may be the reason for higher species richness and diversity in FC system while, in SFC system about 73.9% ( see Table [5]) responded that clearance under coffee were twice. Thus, there were repeated removals of undergrowth vegetation to improve coffee productions consequently lower diversity of bee flora here in this finding. According to Senbeta and Denich,( 2006) higher value of Shannon diversity indices (H') in FC area, and noted that FC have more species and important for the conservation of important plant genetic resource including wild coffee species. Furthermore, altitudinal differences of FC and SFC may cause also in species difference. Moreover, the main problems regarding in honey production in management systems are summarized in the following table.  Kumsa, 2016;Gebretsadik and Negash, 2016), that revealed that developing countries face different constraints in beekeeping sub-sector such as; races of honey bees, honey bee diseases, predators and parasites, the loss of bee floras. Moreover, the similarity coefficient between the sites shows low similarity values. (Jaccard's similarity coefficient = 0.29) (see Table [ 4]). This is that the both forest management systems had low in similarity of bee floral species compositions for both forest management systems.
Likewise bee flora compositions were varied with the source forest management types(X 2 = 1747.289, p <0.001) and between the two management systems (see Table [ 4]).There were significantly different from observed and expected value related to bee flora sources of forest management systems. Each of these types of relationship involves some form of differences between the observed and expected values. This indicates that FC provides more bee flora than SFC management systems or the conversions of FC activities arose negative Journal of Biology, Agriculture and Healthcare www.iiste.org ISSN 2224-3208 (Paper) ISSN 2225-093X (Online) Vol.10, No.16, 2020 14 significant affect on bee flora compositions.

Features of forest beekeepers and trends of honey productions of the district
From the survey data result the average numbers of traditional beehives owned for individual HHs from last five years were ranged from 47.8-62.72 for FC and 44.84-24.36 (see fig. [3]) for SFC management systems. The aim of this survey was to state the situation of traditional honey production system with increasing of the SFC management intensities.

Figure 3.Numbers of average traditional hives per beekeepers of both sites
From the results, the average traditional hives of HHs of FC system is increasing trend where as it was highly decreased in SFC. Even if the hive hanging may has a risk of falling on tree cause death, the preference of SFC system for honey bee keeping became decreasing and the problem may be due to FC conversions, that is why the farmers were focusing on coffee cultivations. Furthermore, trees and lianas that are used in traditional hive making were highly decreased. This in turn causes the decrease in bee colonies and honey production from time to time. Thus, most of the respondents reasoned out that pests and predators, migrations and absconding followed by FC land conversion and high level of its management intensity are responsible for the observed decreases (see Table  [ 5]).On the other hand, about 95 % (see Table [5]) responded that honey trunk or 'holka' disappeared consequently there were no addition of a new swarm of bees to forest. This also agree with (JICA 2005) who states that honey productivity in SW has been reportedly declined due to deforestation, forest conversion and trunk honey or 'Holka' disappeared. This finding revealed that the trends of honey productions and bee colonies were decreasing from time to time in both forest types but for SFC system the situations are more observed (see fig. [4]).

Figure 4. Trends of honey productions
From the results it can be seen that the total average annual honey productions(very valuable forest derived plant crop products) per beekeepers were in a decreasing trend in both sites, except 2006 E.C. production year for FC site that showed higher which was73.9kg/beekeeper ( fig.4).This results are greater than previously reported in this area by JICA, (2005) which states that about 34 kg of honey may be produced in a high production year by one HH for that of beekeepers at FC management system but, less results were recorded for beekeepers at SFC management system which was (32.2-15.2)kg (see fig. [4]) annually for that of the last two years than previously reported one, in this finding. Thus, continued observed declining in honey productions of SFC management systems may be due to the forest is becoming less preferable for beekeeping activities, which may accompanied with the intensity of SFC management systems that resulted declining of bee flora resources. 3.3.2. Current honey production status Numerically the honey yield of FC was higher than SFC system in both high and low production years. Thus, the obtained data result showed that honey yield was an average 9.58 Kg/hive for FC and 6.44Kg/hive for SFC in high production year while 6.5 kg/hive for FC and 4.24 Kg/hive for SFC in low production year(see fig. [5]).  Tadesse et al., (2014) who states that land use changes decreased the amount of forest honey production because of loss of native bee forage and beehive supporting tree and shrub species. Findings of present study indicated that the rate of FC conversions is increasing from time to time to improve coffee productions, of which about 76.9% (see Table [6]) of respondents argued that because of coffee price increases and the remaining are due to population increases, shortage of land and followed by ownerships problems. Surprisingly one recent study by Mitiku et al. (2018) revealed that coffee intensification process does not result in improved coffee productivity nor in increased profits and they found that less intensive coffee production and conservation of forest tree species is a viable sustainability strategy even from an economic point of view. So if further coffee intensifications did not add up on economic profitability as the same times decreasing of forest honey production because of FC conversions, therefore it requires urgent conservation measures.  Table [6]) respondents selects and protects those bee forage species. The finding is similar with Hundera (2013) which revealed that farmers protect those plant species either for bee hive installations or a large amount of its flower resources in their coffee forest even if these species are not good for coffee shade. Even if the respondents argued that they protect the bee tree, the result from fieldwork justified that the reductions of important bee forage species in SFC system were very high.

Physicochemical characteristics of honey
The physicochemical properties of honey play an important role in determining the honey quality and can be affected by bee floral origin and its diversities including the purity of environment. Thus, the present study dealt with the major physicochemical properties of honey in relation to whether the conversions of FC to SFC forest management systems and diversity of bee floral differences related to quality aspects as indicated below. Moisture content:-The minimum, maximum and average of moisture contents of the honey from both forest management systems analyzed in the present study were indicated Table [ 8]. The percentage of moisture content of the honey samples obtained from the study area ranged from 18.1-22.3 with a mean value of 20.03.The moisture content of honey from FC site ranges from 19.2 -22.3% with mean 20.5(n=3), while honey sample from SFC site ranges 18.1 -22.3% with mean 19.5(n=3).There was no significant difference (P>0.05) between the both forest management systems in percentage of moisture content of sample honey collected (see Table [ 8]). The finding is in line with Getu and Birhan,( 2014) which states that no significant differences were observed in % of moisture content between honey samples obtained from the different locations in the same study area. However, the mean average moisture content of SFC is slightly lower indicates that good quality. Two samples exceeded (22.3%) the limit allowed by the Codex and Council of the European Union (EU) of <21% from both sites indicates that the honey was low quality interims of % of moisture content. But all average honey samples are within acceptable range (20.03%) which is similar finding with Getu and Birhan,( 2014)   High moisture content increases the probability risk that the honey will ferment upon storage. The final water content of a honey sample depends on a number of environmental factors during production such as weather, humidity amounts inside the hive, nectar conditions and treatment of honey. Here in FC site the distance between 'mume' and their residential area was far away and the harvested honey was transported with horseback(traditional) in addition to this, the environment has very high humidity that contribute to moisture content of honey which directly related to postharvest quality loss. The study result was similar to Brno ( 2007) who conducted research in rural areas of Kenya, which states that the final water contents of honey samples depends on a number of environmental factors such as weather and amount of humidity in the hives. Furthermore Odessa and Penteado, (2011) revealed that the highest moisture content observed in the organic honey was due to the climatic conditions because the air saturation and the big nectar flow that happens after the rains. All the honey samples (n=6) analyzed for moisture had higher moisture content than the acceptable minimum limit, an indication that most farmers harvest ripened capped honey and that generally honey was harvested at peak harvesting season. Ash (mineral) content: -The ash content of the honey samples obtained from the study area ranged from 0.19-0.49g with a mean value of 0.31 (see Table [ 7]). The average ash content of honey collected from FC (0.24) numerically lower than the average ash content of honey samples collected from SFC(0.39) (see Table [8]), there is a significant difference (p<0.05) in ash content between honey samples obtained from the two forest management types. The ash content of all the analyzed honey samples fell within the 0.01-1.2% range reported by the Ethiopian Quality and Standards Authority to (ES 2005) and 0.6% maximum limit reported by the International Honey Commission (Bogdanov 1999) the mineral content of honey. Thus, the mineral content of honey is related to the geographical and botanical origin of the honey. This suggestion stated that ash content of honey depends on the material contained in the pollen. In this finding within the same geographical locations and same harvesting seasons there was difference observed in both forest management systems. According to Hamza M.Abu-Tarboush, (1992) honey normally has low ash content and this depends on the floral type used by bees. This may be related to diversity of bee flora species or system of forest coffee management. This is further supported by Controle et al. (2013) which revealed that ash represents a direct measure of the inorganic residues left after honey carbonization, and this variability in the ash content can be explained by the floral origin of the honey. pH:-The pH of sampled honey as a factor of FC conversion and its interaction is presented in (see Table [8])). From the result the value of honey pH showed that a significant difference (p<0.05) between honey harvested from FC and SFC management systems. Thus, the mean pH values of honey samples from FC were (3.72) which were lower than honey harvested from SFC (4.07).The variation may be due to diversified bee floral sources. This finding is in line with Shahnawaz et al., (2013) which states that floral difference may also cause the ranges of pH. There is a direct relationship between ash contents and pH, having higher ash contents result, higher pH value according to Brand-Miller, (2005) findings which is supported by this study.i.e. honey samples from SFC shows less acidic. Here honey from FC site is better than honey from SFC interims of the result of pH value, which shows more acidic. This indicates that the low pH of honey inhibits the presence and growth of micro-organisms and makes honey compatible with many food products in terms of pH and acidity (Areda 2015). However, both honey samples ranged between 3.65 to 4.36 and an average of 3.89(see Table [ 7]), which is the international acceptable pH value of honey. The finding is similar to that of Gebru, (2015) who conducted research at Eastern Tigry region revealed that the average pH of honey was 3.86.Furthermore the finding is similar to that of honey from Luso region (Portugal) Silva et al. (2009) revealed that the mean pH value of honey was 3.88. Electrical conductivity: -The electrical conductivity (EC) is a good criterion of the botanical origin of honey and thus is very often used in routine honey control. The EC level of honey samples analyzed in the present study, ranged from 0.17 to0.29 milli Siemens per cm (mScm -1 ) with a mean value of 0.223mScm -1 (see Table [7])The mean conductivity of honey samples obtained from SFC(0.253) system was significantly higher(p<0.05) than that obtained from FC(0.193)mScm -1 system(see Table [ 8]). The conductivity depends on the mineral content of the honey; the higher mineral (ash) content, the higher the resulting conductivity. Here in these findings high ash content was recorded in honey from SFC system and conductivity as well. According to De Controle et al.2013 the differences in EC of the various honeys are attributable to their differing geographical and botanical origins; this can serve to characterize different varieties of honey. This is also supported by (Bogdanov and Stefan 1999). In a given geographical area, ash and acidity were useful for determining the botanical origin of honey. The best indicators for discrimination of honey with varying geographical origin were pH and electrical conductivity as a function of changes in the concentration of honey. Free acidity: -The free acidity level of honey samples analyzed in the present study, ranged from17.60 to31.20 meq/kg with a mean value of 24.79 meq/kg see Table [7]).The mean acidity of honey samples obtained from FC (28.53) system was significantly higher (p<0.05) than that obtained from SFC (21.06) system (see Table [ 8]). This variation may be due to difference of organic acids present in honey. Most of organic acids are present in honey in the form of esters, which contributes to its characteristic flavor and aroma. Some of the acids are introduced into honey via the nectar. The variation here in this finding may also be related to the difference of diversity of bee flora resources. This was agree with (Yadata 2014) (Yadata 2014) (Yadata 2014) (Yadata 2014) andCouvillon et al. 2016 which revealed that the acidity of any honey is directly related to the floral sources that created. Thus, the highest average acidity recorded was due to the diversity of honey flora sources. The result supports each other, hence the low pH value (high acidity) recorded in the FC system and high pH value (low acidity) recorded in the SFC system. However , Brno, (2007) reported that the considerable variation in the amount of acids in honeys perhaps reflects the time required for nectar to be completely converted into honey under differing conditions of the environment, colony strength and the sugar concentration of the nectar of floral sources.

Total soluble solids (TSS):-
The results of the TSS are presented in Table [7]. It was ranged from 63.0 to 73.9%with a mean value of 70.35% (brix). Honey from the SFC (71.13) site had somewhat higher TSS%, whereas that from the FC site had the lower percentage of TSS (69.56). In a forest management system there was insignificant (P > 0.05) deference in TSS of total honey samples. Thus, the variation may be related to the botanical origin of honey or diversity of bee flora (see Table [ 8]). According to Amber, (2003) and Bogdanov, (1999) analysis of TSS can yield the valuable information about the floral origins and can also combines fructose and sugar content should not be less than 60g/100gfor blossom honey (see Table [ 8]).Therefore, the analyzed honey is within an acceptable range. In general among the determined parameters i.e. (pH, ash content, electrical conductivity and free acidity) of honey sample from SFC shows decline in quality than honey samples from FC site compared with Harmonized Methods of the International Honey commission and Ethiopian honey quality Standard (ES, 2005;Bogdanov, 2009),while only moisture and TSS contents were not influenced or did not shows significant differences by either bee floral diversity sources or FC conversion consequences. Furthermore, about 29.1 %( see Table [5]) respondents' uses chemical and pesticide application in their SFC management systems while, such activities were not observed in FC. These activities may bring effect on bee flora species and bee communities that may be accompanied with lower honey productions and quality.

Conclusions
The present study provides preliminary information on the effects of coffee management on the diversity of bee flora, honey yield and quality. It clearly indicated that FC conversion and intensification to SFC has negatively affected bee flora diversities. The study designated that the rate of FC conversions is increasing from time to time, and increasing coffee price is as main driving forces. Thus, the increasing management intensity of SFC systems results in lower bee flora compositions and diversity. Furthermore, honey production of SW part in general and Gera district in particular is decreasing trend. This is shown to be related to FC conversion factors and preference of SFC system for honeybee keeping became diminishing. These finding implies that conservation of bee floral species of FC is a viable sustainability strategy from a biodiversity point of view, and that initiating smallholder beekeepers is a feasible activity in the arena of conservations and as well as a key for improving ecological services. Moreover, the intensification of FC conversion activity relates to the declining of honey quality. The biochemical variation in the composition is significantly different (P<0.05) in ash content, pH, EC and free acidity when comparing FC with SFC honey samples while the percentage of moisture and TSS contents were insignificant (p>0.05) which may indicates that both were not influenced by either bee floral diversity sources or FC conversion consequences. Generally, the study revealed that; FC conversion to SFC (i.e. coffee management effects) are associated with declining bee flora diversity, honey yields and on top of this, it has implications for honey quality loss of SFC systems. Thus, there is an urgent need for control and monitoring on the expansion of SFC cultivation, which needs immediate conservation measures. Therefore, conservationists have to find other means of economic incentives for the community to refrain from further FC conversion and intensifications. Further research has to be conducted on the effect of coffee management on honey quality and ecosystem services.
Author Contributions: Shabu Jemal Abakorma: conducted the field work and wrote the manuscript. Prof. Kitessa Hundera, and Mr. Zerihun Kebebew(Asso.Pro): have supported and enhanced the quality and structure of the study and contributed ample of technical, logical and structural supports from the beginning.