Fieldwork in and around Arabuko Sokoke forest

The coastal forests along the East African coastline of Somalia, Kenya and Tanzania hold a large number of endemic species, which are under extreme threat due to anthropogenic habitat transformation. Hence, this coastal forest were classified as one of the global biodiversity hotspots (according Conservation International). The most representative and largest remaining fragment of tropical dry coastal forest is the Arabuko Sokoke forest (ASF) in southern Kenya. Two iconic representative species occurring exclusively in this forest remnant are the Golden-rumped Elephant Shrew (Rhynchocyon chrysopygus) and the Sokoke Scops Owl (Otus iraneae). Despite intensive protection and efforts in conservation management, the ASF is highly threatened due to the exploitation of natural resources which negatively affects habitat quality. Increasing demand for land and natural resources (agricultural land, timber, charcoal, bush meat) have a negative impact on this ecosystem. Reduced habitat quality caused a rapid decline of many species and their local populations, which leads to an increasing risk of local extinctions. Effective conservation programs in this human-dominated area are essential to protect both, intact environments (including ecosystem functions for human livelihood needs) and high-quality habitats (home of many endangered endemic species) – impressions below.

Field courses and training was conducted from 25.2.-17.3.2017. In order to achieve these objectives, teaching, field courses and workshops were carried out within the transdisciplinary DAAD framework “Reconciling human livelihood and nature conservation in East African forest biodiversity hotspots” during spring 2017. Students and scientists were split into smaller groups to work in different thematic fields and on various research questions. These activities were conducted in close collaboration with Pwani University and other universities and stakeholders, like: South Eastern Kenya University, Taita Taveta University, Kenyan Forest Service, Kenyan Wildlife Service, Kenya Forestry Research Institute, National Museums of Kenya, Nature Kenya, A Rocha Kenya, Friends of Arabuko Sokoke, and many others! Field work (i.e. data collection) was conducted in three different areas: (i) outside the forest in the agricultural land, (ii) inside the forest covering different forest types, and (iii) along transects transgressing the forest border and ranging from the intact forest into the degraded agricultural land. Scientists and students worked in Kenyan-German tandems, in the following thematic fields:

Land tenure: Land use and land tenure affect the exploitation of resources of Arabuko Sokoke forest. In this working package, two methodological approaches were combined: Land mapping and interviews with the respective land owners. The properties of single households were mapped with GPS devices i.e. tablets and details on land property and land use were analysed. According to the information collected in the field, most of the land is used for subsistence agriculture (cultivation of maize, cassava, cassaurina, coconut, cashew nut, mango). Main problems arise from periodic droughts and human wildlife conflicts (with baboons and monkeys mainly). The land tenure systems differ strongly and range from freehold (mostly family land), leasehold, public land, squatters, and community land. Thus, tenure insecurity are particularly high if land is not freehold and an own property.


Find below some impressions from the field work (including a gym found on one of the parcels analysed).

Drone based forest disturbance assessment: Detailed aerial pictures provide information on the condition of an ecosystem. The Arabuko Sokoke forest consists mainly of three forest types (Cynometra forest, Brachystegia forest, mixed forest). The forest suffers under anthropogenic exploitation (tree logging or charcoal production), as well as by elephant disturbances. To distinguish among different forest types and to analyse the degree of habitat disturbance we used unmanned aerial vehicles (UAV), here a DJI Phantom 3 drone, equipped with two different cameras (normal RGB and a near-infrared camera). Flight missions were programmed as circles with a 50m radius. For each category we performed 24 overflights. Pictures were taken from 60m altitude. After image acquisition the next steps will be the processing of RGB and near-infrared imagery and afterwards combination of both to compile different vegetation indices.  These data will provide information about the condition of the forest and give us insight into disturbances from human and elephant activities. Find below some impressions during the collection of data.

Social science: Most strategies in nature conservation (see also various textbooks) are based on pure ecological i.e. biological data, mostly neglecting the human dimension like livelihood needs, attitudes and culture. Furthermore, these strategies also do not incorporate national and/or local governance structures or cultural behaviour. However, knowledge about human livelihood needs is particularly important in regions characterised by strong demographic pressure and the degradation of natural resources. Such transdisciplinary approaches may provide much more realistic strategies. More than 250 questionnaires were completed and expert interviews were conducted in three settlement areas around the Arabuko Sokoke forest (see below).


Main goal of these surveys were to gather information about socio-demographic structures, land tenure and land use, awareness and attitude, willingness and involvement. Main target group were the local farmers conducting subsistence agriculture around Arabuko Sokoke forest. Interviews were held in Kiswahili. First results indicate that people live at low income level (<50€ per month) and have a low level of education. Most of the respondents wish to benefit more from the forest and the activities around it (i.e. on alternative income sources) and would like to be stronger involved in policy making and decision taking. Some impressions from the field work see below.

Rapid ecosystem function assessment: Ecosystem functions are crucial for life quality of people, and are assumed to significantly decrease with habitat disturbances and habitat deterioration. In this working package, we raised and elaborated the following research question: Do ecosystem functions differ between the intact forest and the disturbed and partially degraded agricultural land? Three parameters were assessed using simple, cheap but meaningful parameters (according the REFA method: Meyer et al. 2015) pollination, predation, and seed dispersal. These factors were analysed along 10 transects covering the forest and the agricultural land (with 9 locations for each transect, 2 inside the forest, 1 at the forest edge, 6 in the agricultural land). Details on the data collected see below:


First results indicate a higher level of pollination, predation and seed dispersal inside the forest than in the agricultural land. Spill over effects were strongest close to the forest and decreased with the distance to the forest. Apart from students and local assistants, also pupils of neighbouring schools were involved in this activity (see pictures below).

Vegetation ecology: Originally, the entire East African coastline was covered with coastal forest. However, this vegetation type became almost entirely cleared and transformed into agricultural land and exotic tree plantations. We assume, that some few typical forest species are still left as a kind of `legacy´ across the agricultural land. Furthermore we expect that typical forest species can be found across fallow land in the wake of regeneration and reforestation. In parallel, we assume that negative edge effects along the forest border might negatively impact the pristine forest. Trees and shrubs with >5cm DBH (diameter at breast height) were assessed along transects ranging from the agricultural land into the forest (see pictures below). These data will provide insights about the degree of disturbances in the forest (mainly through anthropogenic activities), and information on the vegetation structure beyond the forest.

Invertebrates in canopies: Most species in forest habitats are found in the canopies of trees. Thus, collecting arthropods in canopies is crucial to learn more about biodiversity sensu lato (species, abundance, distribution). We used flight interception traps equipped with LED lights. Traps were set in the following four habitat categories: Cynometra forest, Brachystegia forest, mixed forest, and forest border (transgression from forest into agricultural land). Each habitat category was represented by three locations, each location was sampled during 4 nights. LEDs were turned on from late afternoon until late morning and traps were activated during this time. Insects were stored afterwards in ethanol. Representatives of the following groups were collected: Cicadina, Aphidina, Heteroptera, Coleoptera, Neuroptera, Apocrita, Formicidae, Lepidoptera, Nematocera, Brachycera, and others. These data will provide information about differences in species composition and abundances among sampling sites i.e. habitat types and allow to evaluate the impact of negative edge effects. This activity will be continued during the next months which will allow a seasonal comparison (dry season vs. rainy season) afterwards. See below a sampling site in the Cynometra forest.

Butterfly communities: Species communities frequently reflect habitat structure and habitat quality. Analysing species communities provide valuable information about potential edge effects and spill over effects and provide valuable information about the role of exotic tree plantations acting as surrogate habitats. Butterflies are known to react highly sensitive on environmental changes and strongly depend on intact habitats. Here we counted butterfly species (taxa, abundances) along 150m transects, set in the following habitat types (pristine and disturbed i.e. artificial): Cynometra forest, Brachystegia forest, mixed forest, forest edge and plantation of exotic trees. For each habitat category we worked along 10 transects. Each of these transects were visited 10 times during a period of 3 weeks at the end of the dry season. Transects were ecologically characterised according soil, vegetation (availability of nectar sources), and structural parameters (forest density, shade). For each taxa assessed we assigned ecological parameters (habitat specialization, distribution, behaviour, according to Larsen 1996). Transect counts along the same transects were already conducted by Sina Hesse (a master student) during the rainy season in 2016. This joint data-set will provide information on how species composition differ among habitat type (including pristine vs disturbed), and how seasonal conditions like the availability of nectar sources affect the species composition (i.e. the distribution of butterflies across the forest). See below six representatives of the >60 butterfly species assessed.

Small mammals: In this working package we compare species composition of small mammals along a gradient from the forest into the agricultural land, to test for spill over effects and potential edge effects. Small mammals were collected using Sherman live traps. At each of the 8 transects, 23 Sherman live traps and 8 pitfall traps were activated. Data on species diversity, abundance of taxa and information on body condition were collected. Traps were baited during the evening and controlled (and closed) during the morning after sunrise. Traps were deactivated during the day. The caught individuals were identified, sexed, the reproduction status was analysed and each individual was marked. The following measurements were taken: weight, tail, body length, foot length. An ear tissue was taken to analyse the health status. Afterwards, individuals were released at the same place. In addition, environmental parameters (vegetation, structures) were also assessed. As the collection of individuals during the dry season was difficult, this working package will be continued during the next month. This extended data sets may provide information about differences in species composition, abundance and health status across seasons (the dry vs. rainy season). See below the small mammal team and an activated Sherman live trap.

Tropical field ornithology: Organisms suffer under stress particularly in disturbed environments. In this working package we raise the following research hypotheses: (i) bird species assembly differ between forest and agricultural land; (ii) body conditions are better inside (or close to) the forest than in the agricultural land; (iii) habitat specialists react more sensitive on habitat disturbances than habitat generalists. The team collected birds along 6 forest-agricultural land-transects, with 7 sampling points of each transect (2 in agricultural land, 4 inside the forest, 1 at the forest edge). Sampling of birds was conducted with mist nets (2 mist nets per sampling point – in total 14 nets). Nets were opened from 6am to 10am to catch bird species (habitat generalists as well as habitat generalists). The following parameters were assessed: physiological measurements, reproduction status, body condition. Blood was taken to analyse the H/L  ratio (ratio of heterophilis to lymphocytes in the blood). The team and some impressions taken during early bird collection are shown below.

The team: The following people participated during field work: Scientists – Anna Schopf, Sebastian Meyer, Christina Fischer, Roland Gerstmeier, Marco Rieckmann, Thomas Schmitt, Jan C. Habel, Mike Teucher, Kim Geraldine Mortega, Gesine Heinrich, Tobias Bendzko, Christine Schmitt; Students: Nadine Bihler, Stefan Reuter, Ina Zeschmann, Elena Hülsemann, Julia Weißhaupt, Anja Wiese, Janika Kerner, Elisabeth Koc, Jose Rangel, Julia Gruber, Claudia Dieckow, Leonie Schmid, Lukas Bofinger, Michaela Unterbichler, Anna-Katharina Zech, Simon Charo Jumaa, Hamisi Tsama Mkuzi, Vincent Okelo Wanga, Ruth Lozi Muranga, Noreen Mutoro, Beth Njeri Muhiu, Timotih Musa, Bonventure Obeka; Field assistants: Onesmus Kioko, Kennedy Matheka, Morris N. Mutua, David Ngala, Francis Kahinio, Charo. See our group picture below (with most of the group members).


This activity was kindly supported by various stakeholders and people: Kenya Forest Service (here mainly Blessingtone Maghanga) gave us allowance to work inside the forest (together with elephants and buffalos), the Kenyan Wildlife Service (here mainly Lynn Njeri), Kenyan Forest Research Institute, A Rocha Kenya, Nature Kenya, Friends of Arabuko Sokoke (here mainly Ann Robertson and Dr. Maarifa Mwakumanya), and David Ngala! And our very reliable colleagues and main collaborators Prof. Halimu Shauri and Dr. Maarifa Mwakumanya from Pwani University. Some of the field work activities will be continued during 2017 by our local collaborators, Kenyan and German students. See below some of our collaborators – David Ngala (left), students and scientist from SEKU, Lynn Njeri from Kenyan Wildlife Service, and very supportive field assistant Onesmus Kioko (together with Fabian).

Field work in and around Arabuko Sokoke forest was exceptional! The nature and wildlife in the forest, but also in the nearby Mangrove forest is overwhelming. The stay in Jamii Villas was the best choice, to be close to the forest, together with the local people, and in a very nice surrounding with excellent logistics. Apart from our field work we enjoyed night walks to see elephants and the endemic endangered Sokoke Scopes owl, visits to Mida Creek, Gede Ruins, or to the beach of the Indian ocean.