Study Area

The study was conducted in the lower River Tana basin in Tana River County, Kenya (latitudes 0^o0’53”30” and 2^o 0’41” South, longitudes 38^o30’ and 40^o15’ East; Figure 1). The basin hosts crocodile egg harvesting zones belonging to three privately-owned crocodile ranches: Galaxy Croc Farm, Kazuri London Limited and Nile Crocodiles Limited (Galaxy, Kazuri and Nile, respectively, hereafter; Figure 1). While the primary facilities of these ranches are located outside Tana River County, they operate seasonal crocodile egg harvesting and incubation field units adjacent to the river.

The annual rainfall in the study area ranges between 400mm and 750mm, and generally occurs in March–May (“long rains”) and October–December (“short rains”). However, rainfall exhibits high spatial and temporal variability and the region is prone to frequent droughts and flooding (Odhengo et al., 2014; Mohamed, 2015). The mean annual temperature ranges from 23^o to 33^o Celsius (Odhengo et al., 2014). Notably, the region is home to many endemic and range-restricted species including the Tana River red colobus monkey (Piliocolobus rufomitratus) and the Tana River crested mangabey (Cercocebus galeritus) (Odhengo et al., 2014). Additionally, the region hosts the Tana River Primate National Reserve (TRPNR), a government-protected conservation area that serves as an important refuge for many animal and plant species.

Tana River County has a human population of 314,710 (KNBS, 2019). Due to generally dry weather conditions across the county, human settlements are mainly concentrated in areas adjacent to the river (Mohamed, 2015). The county is mainly inhabited by the Pokomo, Orma, Wardei, Watta and Ilwana (also known as Malakote) ethnic groups. The Orma, Wardei, Watta, Ilwana, and Pokomo occupy the upper part of the river and are largely Muslims, whereas the Pokomo occupy the lower part of the river and are mainly Christians (Githui, 2020; Joshua Project, 2019; Ylvisaker, 1982).

Figure 1.

Map of the study area showing the different egg collection zones and location of the crocodile ranches.

The Pokomo practice farming, fishing and sand harvesting, while the Wardei and Orma combine livestock keeping with some shifting cultivation (GoK, 2018a; Mohamed, 2015; Terer et al., 2004). The Watta, originally hunters and gatherers, have now transitioned to harvesting and selling forest wood products and engaging in small-scale farming (National Gender and Equality Commission [NGEC], 2014). The Ilwana community practice farming, beekeeping and livestock keeping (GoK, 2018b). Other ethnic groups in the region include the Munyoyaya who practice farming and beekeeping, and the nomadic-pastoral Somalis (from the neighboring Garissa County) who utilize the river basin for pasture and water during dry periods (Food and Agriculture Organization [FAO], 2017; Mohamed 2015).

Target Population

The target population comprised households in villages located within 2km and 5km of the river bank at the non-delta segment and delta segments of the river, respectively. These villages were selected because of their proximity to the river and the associated high likelihood of their residents to interact with crocodiles. The selected villages contained a total of 4,363 households, representing approximately 12% of the total number (35,099) of households in the entire Tana River County (KNBS, 2019).

Sampling Procedure and Data Collection

We conducted the study from February to March 2019, using a combination of quantitative and qualitative data collection techniques. Specifically, we conducted a household survey, seven key informant interviews (KIIs) and four focus group discussions (FGDs). The household survey involved 365 households randomly selected from ten villages situated adjacent to the river. Four, two and four of these villages were located within the egg collection zones for Galaxy, Kazuri and Nile, respectively ( Supplemental 1). The overall sample size was proportionately distributed among the villages based on number of households. To obtain a random sample from each village, we used a household list prepared with the help of the village elder.

We conducted the household survey through face-to-face administration of a semi-structured questionnaire to each of the selected household heads or their adult representatives. The questionnaire comprised two parts; the first part elicited information on the respondents’ sociodemographic characteristics, while the second part focused on their crocodile utilization practices. Crocodile utilization was categorized into three: “legal”, “illegal” and “non-utilization”. A participant was considered to be involved in legal practices if he or she reported engagement in one or more crocodile utilization activities not prohibited by Kenyan law. These activities included collecting crocodile eggs for ranches, locating crocodile egg nests for ranches, incubating crocodile eggs and selling hatchlings to ranches, training egg collectors and working at a ranch or field incubation unit. Conversely, a participant was deemed to be involved in illegal practices if he or she indicated engagement in one or more activities prohibited by the law such as eating crocodile eggs obtained from the wild, eating wild crocodile meat, hunting crocodiles, using crocodile meat as fishing bait, selling crocodile eggs to villagers, selling crocodile meat to villagers and selling crocodile meat to fishermen. Non-utilization referred to non-involvement in either legal or illegal crocodile utilization practices.

We conducted KIIs and FGDs using interview schedules and prompts. The key informants were managers of the three crocodile ranches, and relevant Kenya Wildlife Service (KWS) and Tana River County officials. The FGDs consisted of six to eight discussants each. Two of the four FGDs consisted of household heads (both men and women), whereas the rest consisted of egg collectors, all of whom were men. For in-depth focus, the KII and FGD participants were selected on the basis of their knowledge, experience and role in crocodile ranching and conservation in the study landscape. The themes covered in the KIIs and FGDs included local community benefits from ranching-based crocodile utilization, crocodile utilization beliefs and practices, human activities along the river, and potential anthropogenic threats to crocodiles.

Data Analysis

Quantitative data were analyzed using the Statistical Package for Social Sciences (SPSS) computer program version 20 (International Business Machines Corporation [IBM], 2011). We summarized these data using descriptive statistics, particularly frequencies and percentages. In addition, we fitted a multinomial logistic regression model to investigate the influence of various sociodemographic factors on participants’ involvement in different crocodile utilization practices. The dependent variable contained three levels of crocodile utilization, namely, “legal”, “illegal” and “non-utilization”, with the latter serving as the reference category. The sociodemographic factors included in the model were age, level of formal education, annual income, gender, ethnic group, religion, main source of livelihood and egg collection zone. The factors age, income and education level were used as covariates (after categorization and ordering of their respective categories), whereas all other factors were used as nominal variables. For each explanatory variable, the selected reference category was the category in which the lowest (or lower) proportion of participants reported involvement in illegal utilization. The reference categories for the different explanatory variables are presented in Table 1.

We analyzed qualitative data thematically using the Max Software for Qualitative Data Analysis (MAXQDA), in accordance with (Kuckartz & Stefan, 2019). The analysis entailed looking for similarities and differences between the texts and summarizing the results. To achieve this, we first organized the data into two documents (KIIs and FGDs) and then conducted a systematic study of the files to understand the whole text. Finally, we categorized contents into codes (themes), assigning a unique color to each code.

Sociodemographic Characteristics of Household Survey Participants

Most (68%) of the household survey participants were male. The age categories 30-39 and 40-49 years were the most frequent, each accounting for 23% of the participants. A majority (61%) of the participants were Muslims (Table 2 and  Supplemental 2). Pokomo was the most dominant ethnic group, accounting for 63% of the participants (Table 2 and  Supplemental 2). Slightly less than half (47%) of the participants had attained primary level of formal education, while less than 10% of had tertiary education. Most (49%) of the participants indicated that farming was their main source of income. A majority (70%) participants had annual income levels lower than US dollars 2,336 (Table 2 and  Supplemental 2).

Table 1.

Reference categories for explanatory variables.

Table 2.

Household participants’ sociodemographic characteristics by crocodile use category.

Figure 2.

An illustration of human activities harmful to crocodiles along lower River Tana. In (a) and (b) an island on the river now encroached for crop cultivation but hitherto considered a refuge for crocodile breeding; (c) sand harvesting at the river; (d) poison wrapped in goat skin at the riverbank targetted at crocodiles; (e) Aloe ruspoliana plant, locally called “rasai” used to poison crocodiles; and (f) crocodile nest destroyed by machettes. Photos by Mosse.

Anthropogenic Threats to Crocodiles

The KII and FGD participants highlighted multiple human-induced threats to crocodiles (Figure 2a-f). One frequently mentioned threat was expansion of agriculture into the riparian zone, which is an important crocodile nesting habitat. Although the participants indicated that the condition of crocodile habitat was still healthy in some areas like the Tana Delta, they expressed fear that this situation would not last for long because of continued expansion of crop cultivation into the riparian zone, which results in land clearance and hence the destruction of this critical crocodile habitat.

Another frequently mentioned threat was retaliatory killing of crocodiles and the destruction of their eggs as a result of human-crocodile conflicts. In particular, it was revealed that Wardei and Somali livestock keepers, especially in the upper segment of the river basin, often take such actions to retaliate against livestock losses caused by crocodiles. Conversely, it was revealed that Orma livestock keepers were more tolerant to crocodiles because of their belief that their god (“Allah”) protects people and livestock from crocodile attacks. It was also reported that most local community members do not benefit from the existing crocodile ranching program. They also indicated that, contrary to the provisions of Kenya’s Wildlife Conservation and Management Act, 2013 (GoK, 2013), compensation for crocodile-related livestock losses is rarely provided, which further exacerbates human-crocodile conflict in the region. Notably, participants highlighted that due to these factors, local community members continued to engage in various practices that are harmful to crocodiles. For instance, some of the discussants revealed that crocodiles are commonly poisoned using Aloe ruspoliana plant (locally called “rasai’). Specifically, they reported that the leaves of this plant are usually cut into small pieces and wrapped in animal skin to form a bait, which is then strategically placed in areas frequented by crocodiles such as their basking sites.

The participants also identified consumption of wild crocodile meat and eggs by some local community members as a major threat to crocodiles in the region. Furthermore, they indicated that the number of eggs available for harvesting by crocodile ranches continues to decline partly due to these practices. This view was captured by one crocodile ranch manager who stated, “I now have less and less to do with the Pokomo because there are no eggs available for collection in their areas”. The participants held the general opinion that the number of breeding crocodiles had declined over the years. One of the key informants asserted, “In the year 2005, it was possible for one egg collector to get 2000 eggs per season but now even an intense search may not yield 300 eggs”. Another threat mentioned was the engagement of some community members in practices such as capturing wild crocodiles using fishing rods, killing them and using their meat as fishing bait.

Level and Sociodemographic Determinants of Crocodile Utilization

Overall, 5% and 32% of the household survey participants reported involvement in legal and illegal crocodile utilization, respectively; the remaining (63%) reported non-involvement. Limited involvement was observed in various forms of legal crocodile utilization, with no more than 3% of the participants reporting engagement in harvesting crocodile eggs for ranches, locating nests for ranches or working at field incubation units (Table 3). Eating wild crocodile meat and eggs were the most common illegal utilization practices, with 31% and 27% of the participants indicating engagement in these activities, respectively (Table 3). However, the other illegal utilization practices were far much less common, with less than 3% of participants reporting engagement in each of them (Table 3).

Table 3.

Number of households engaged in various categories of crocodile utilization practices.

The model indicates that age, annual income, gender, ethnic group, religion and main source of livelihood significantly shaped (χ²>9.359, df=1, p≤0.018) participants’ involvement in different crocodile utilization practices but not (χ²>2.326, df=1, p>0.285) education and egg collection zone (Table 4). Specifically, increasing age increased the likelihood of engaging in illegal crocodile utilization practices by a factor of 1.5, whereas it decreased the likelihood of engaging in legal practices by one-half (both χ²>6.433, df=1, p<0.011; Table 5). Increasing annual income doubled (χ²=8.052, df=1, p=0.005) the likelihood of engaging in illegal practices but did not significantly alter (χ²=0.428, df=1, p=0.513) the likelihood of engaging in legal practices (Table 5). Key informants and focus group discussants revealed that, compared to younger people, older individuals tend to engage more in illegal crocodile utilization activities because of their greater traditional knowledge of crocodile utilization and greater willingness to take risks associated with such practices. They also indicated that younger people tend to have more socio-economic opportunities away from the area where they become more exposed to external influence, thereby making them less likely to participate in illegal crocodile utilization activities. Finally, they opined that as the need for higher income to take care of increasing family responsibilities increases with age, older people are more likely to engage in illegal utilization practices.

Table 4.

Likelihood Ratio Tests.

Table 5.

Effects of various sociodemographic factors on the likelihood of involvement in different categories of crocodile use.

With regard to gender, a male participant was 12 and three times more likely (χ²=4.870, df=1, p=0.027 and χ²=8.698, df=1, p=0.003) than a female participant to be involved in legal and illegal crocodile utilization, respectively (Table 5). The KII and FGD participants revealed that activities such as harvesting crocodile eggs from the wild and hunting crocodiles are deemed risky and are hence largely performed by men. In particular, they revealed that men generally tend to be more willing than women to risk being attacked by crocodiles and other wild animals while travelling on water by canoes in search of crocodile nests. In addition, the participants disclosed that all the canoes used in searching for crocodile nests in the region are exclusively owned and operated by men.

Regarding religion, a Christian participant was 15 times more likely than a Muslim participant (χ²=23.405, df=1, p<0.001) to be involved in illegal (but not legal [χ²< 0.425, df=1, p=0.515]) crocodile utilization (Table 5). The KII and FGD participants revealed that Muslims are prohibited from practices that are harmful to crocodiles. A Muslim FGD discussant from Sala summarized this view by stating, “According to Islam, killing crocodiles is “haram” (proscribed) and therefore we are neither allowed to kill them nor eat their meat and eggs”. With regard to the role of main source of livelihood, crop farmers were eight times more likely (χ²=4.694, df=1, p=0.030) than livestock keepers to be involved in legal (but not illegal [χ²< 0.1, df=1, p=0.996]) crocodile utilization (Table 5).

Although the overall effect of ethnic group was statistically significant, a comparison of the Orma (reference group) against each of the other ethnic groups did not show any significant difference (χ²=1.767, df=1, p>0.184; Table 5). However, KII and FGD participants indicated that the traditional indigenous knowledge, beliefs and cultural practices of the Orma promote human-crocodile coexistence. They also revealed that the Orma believe that their prosperity is associated with the river and its animals and, therefore, play a role in protecting the river ecosystem. Furthermore, they indicated that greater coexistence between the Orma and crocodiles is demonstrated by the presence of undisturbed crocodile habitat and high abundance of these reptiles and their nests in areas occupied by this ethnic group. Conversely, it was revealed that practices such as hunting crocodiles and eating their meat and their eggs were particularly common among a section of the Pokomo ethnic group called “Mila Chini”. The KII and FGD participants further revealed that the “Mila Chini” Pokomos engage in these practices for nutritional, medicinal and cultural purposes. One egg collector from Galaxy stated, “Not all Pokomos eat crocodile meat and eggs. It is the “Mila Chini” Pokomos who live in Garsen and downstream areas of the river who eat these products”.

Anthropogenic threats to crocodiles

The observed multiple anthropogenic threats to crocodiles in the lower River Tana basin, where a ranching program has been in place for decades, suggests that human-crocodile conflict still persists in this landscape, contrary to a recent report (CITES, 2019). Availability of suitable nesting sites is known to be important for ensuring great reproductive success of crocodiles and persistence of their populations (Calverley & Downs, 2017; Refsnider, 2016; Somaweera & Shine, 2013). Therefore, the revealed continued expansion of agricultural activities into the riparian zone of our study region is of major concern because such activities can reduce the quality and quantity of habitat available to crocodiles for nesting, thereby lowering their reproductive performance and destabilizing their population (Somaweera et al., 2019; Utete, 2021).

Our finding that key informants and discussants associated increased consumption of wild crocodile meat and eggs with a decline in the number of crocodile eggs available for ranching is consistent with other studies indicating that such practices may be detrimental to the survival and sustainable utilization of crocodiles (Kahler et al., 2013; Lindsey et al., 2013). Further, the revealed retaliatory killing of crocodiles and destruction of their nests by livestock keepers can undermine sustainable utilization and conservation of crocodiles (Raley, 2016; Treves & Bruskotter, 2014). Specifically, such practices can suppress crocodile populations by decreasing the number of breeding adults and impairing net recruitment, thereby decreasing the quantity of eggs available for ranching operations.

Our finding that failure to financially compensate local community members for crocodile-induced livestock losses aggravated retaliatory killing of crocodiles agrees with other studies indicating that effective compensation schemes can help mitigate human-wildlife conflict (Hazzah et al., 2014; Treves & Bruskotter, 2014). However, such schemes should be implemented with caution in multicultural landscapes such as the lower River Tana basin where, due to differences in cultural beliefs and practices; some communities (e.g. Orma) are more tolerant to crocodiles than others. Specifically, blanket implementation of compensation without being cognizant of the roles played by different communities in tolerating and promoting crocodile conservation can result in inequitable distribution of compensation funds among communities, thereby breeding inter-community conflicts (Nyhus et al., 2005). Therefore, there is need to increase focus on promoting human-crocodile coexistence by adopting preventive measures such as constructing crocodile barriers at livestock watering points and piping water from the river to villages for domestic use (Frank, 2016; Than, et al., 2020; van der Ploeg et al., 2011). In addition, socio-economic incentives such as community projects should be provided to increase local communities’ level of tolerance to crocodiles. Finally, there is need to promote livestock husbandry practices that minimize livestock depredation by crocodiles.

Level and Sociodemographic Determinants of Crocodile Utilization Practices

Our study suggests that the local communities in lower River Tana have largely not been successfully incentivized to derive livelihood benefits from the existing long-term crocodile ranching program. The observed positive association between age and involvement in illegal practices of crocodile utilization is consistent with previous studies elsewhere (Akinsorotan et al., 2020; Lopez-Feldman, 2014; Mogomotsi et al., 2020). This pattern can be explained by our qualitative data indicating that older people have more knowledge of traditional crocodile utilization practices, greater family responsibilities and greater willingness to take risks, thereby making them more predisposed to engage in illegal practices. Notably, at face value, our finding that younger people were less involved in illegal practices suggests that younger people are less focused on hunting and fishing activities. However, our finding that younger people were more involved in legal practices suggests that younger people are more likely to embrace practices that are compatible with crocodile ranching. These findings are indicative of potentially positive prospects for sustainable utilization and conservation of crocodiles in such landscapes.

Previous studies have reported that people with low income levels can be highly dependent on natural resources for their survival, making them more predisposed to engaging in illegal extraction of such resources (Harrison et al., 2015; Kümpel et al., 2010; MacMillan & Nguyen, 2014). Conversely, however, the observed positive association between income level and engagement in illegal crocodile utilization practices suggests that being relatively poor does not necessarily increase overdependence on such practices. The observed gender disparity in engagement in crocodile utilization can be explained by our qualitative data indicating that men exclusively own the requisite equipment (canoes). This finding is consistent with other studies indicating that women tend to be underrepresented in consumptive wildlife utilization activities (Akinsorotan et al., 2020; Anthony et al., 2004; Kümpel et al., 2010; Loibooki et al., 2002; Sunderland et al., 2014).

Our qualitative findings indicated that the Orma have cultural beliefs and practices that protect crocodiles, whereas the Pokomo traditionally consider practices such as eating wild crocodile meat and eggs as normal, despite the fact that the law prohibits them. Consistent with the latter finding, the Pokomo are known to attach great nutritional, medicinal and cultural identity values to eating wild crocodile meat and eggs (Pooley, 2016). This finding underscores the significant role of cultural beliefs and practices in influencing crocodile utilization practices in such landscapes, thus aligning with previous studies (Cuni-Sanchez et al., 2016; Kiffner et al., 2015).

The observed disparity between Christians and Muslims in engagement in illegal crocodile utilization underscores the role of religion in influencing natural resource use patterns among communities (Baker et al., 2014; Brackhane et al., 2019; Reuter et al., 2018). We attribute this disparity to religious differences in practices and norms regarding crocodile utilization, and particularly to the revelation from our KII and FGD that Islamic religion (unlike Christianity) prohibits its followers from engaging in illegal crocodile utilization. Thus, our finding suggests that religious beliefs, practices and norms can play a role in deterring people from engaging in practices that are harmful to crocodiles. Livelihood source has been reported as an important factor influencing people’s natural resource conservation and utilization decisions (Akinsorotan et al., 2020; Kiffner et al., 2015; Li et al., 2016). The observed disparity between livestock keepers and farmers in involvement in legal crocodile utilization is interesting because livestock keepers bear the brunt of coexisting with crocodiles yet they are less likely to participate in and benefit from such utilization. We posit that livestock keepers possibly do not have sufficient time to engage in time-consuming crocodile utilization tasks (e.g., searching for nests) since they spend much of their time looking after livestock.