Exploring crop maturity times as a conservation tool for improving resilience to human-elephant conflict: elephant crop-raiding

ABSTRACT

The purpose of this study was to alleviate negative impacts of elephant crop-raiding faced by smallholder farmers around the Okavango delta. Although many studies have identified the relationship between crop-raiding and climatic factors, little attention was paid to aligning cropping duration to high rainfall. Therefore, this study aimed to investigate how crop maturity duration influences farmers’ vulnerability to elephant crop raiding and food loss. An experimental plot consisting of early and late maturing varieties of maize, sorghum and cowpea was set up and monitored in farms adjacent to corridors. Early maturing varieties yielded more (mean = 0.35 tha⁻¹) than later maturing varieties (mean = 0.25 tha⁻¹). The prolonged stay of crops in fields exposed them more to elephants (p < 0.001). Unlike late varieties, early varieties served as an evasion strategy by allowing earlier harvesting when there was adequate rainfall and elephants rarely frequented farms. Late maturing varieties were harvested at 16–19 weeks when elephants frequented the fields. This harvest timing significantly influences the vulnerability of crops to debilitating elephant raiding. This outcome suggests that policy reviews should incorporate more climate action into smallholder farming systems to strengthen SDG 1: no poverty and SDG 2: zero hunger, ease hostility toward elephants and improve elephant conservation.

KEYWORDS:

• Co-existence
• human elephant conflict
• vulnerability

SUSTAINABLE DEVELOPMENT GOALS:

• SDG 1: No poverty
• SDG 2: Zero hunger

Introduction

Climate change and global warming, in general, have brought about devastating direct and indirect impacts on agriculture in Botswana. These changes have led to notable changes in rainfall patterns and variations in water availability. Rainfall has become more unpredictable, while rainy and cropping seasons have also drastically shortened (Jackson et al. 2008). Crop production by subsistence farmers has been one area of agriculture that has been hard-hit by uncertain climatic conditions (Müller et al. 2011), such as unpredictable rainfall and temperature. These negative changes occur at an exponential rate, leading to the inability of smallholder farmers and cropping systems to adapt. At the same time, phenological development of crops, cropping seasons, agro-ecological zones and population dynamics of living organisms are changing (Allara et al. 2012) as a response to climatic change. The variations in water availability resulting from short rain periods and prolonged droughts often exacerbate the population dynamics of pests and invasive alien species (Allara et al. 2012). These variations bring significant worries, especially about food production and food security.

Farmers in the eastern Okavango Panhandle mostly depend on Government subsidies for inputs into arable agriculture. Mostly, such inputs include seeds, fertilizers and mechanical assistance dispensed through the Integrated Support Programme for Arable Agriculture Development (ISPAAD) programme. The programme’s main goals are to increase food security at the national and subsistence levels, introduce mechanization in farms, and promote awareness of good farming methods (Kashe et al. 2023). In addition, ISPAAD was a tool for Botswana to amplify its efforts toward socio-economic improvement of communities by achieving Sustainable Development Goals, SDG 1: zero hunger and SDG 2: no poverty. In order of preference, the ISPAAD programme dispensed maize, sorghum, millet and cowpeas to these farmers (Marumo et al. 2014). The program was beneficial to many farmers in other parts of Botswana, however, in the eastern Okavango area, cereal crops such as maize and sorghum are very prone to elephant raids.

A combination of climatic factors and elephant crop raiding has been demonstrated in several studies alongside proposals for mitigations. Mosojane (2004), Masunga (2007), Tiller et al. (2021), and LaDue et al. (2021) also reported that rainfall and water availability influenced elephant movements and crop-raiding activities. Although other animals are also reportedly involved, elephants are mostly implicated in much of the crop losses around the Okavango farming communities in Botswana (Songhurst 2017). Osborn (2004) has reported that most crop raids are concentrated in the later parts of the wet season and beginning of the dry season periods. Similarly, Parker and Osborn (2001) reported that most crop raids occur around March to April when many crops reach the ripening stage. The findings concur with reports by Songhurst (2017). The latter also observed no raids during the seedling stage and middle growth stages, but only when the crops were mature, just like the findings by Jackson et al. (2008).

In view of elephant movements being influenced by water availability. Short cycle crop varieties are potentially a promising solution to such unpredictable and adverse changes brought about by climate change and variations in water availability. For example, in Cameroon, maize, sorghum, and millet produced reduced yields due to short rain periods and extended droughts. However, breeding early maturing varieties of these crops could overcome the problem (Guei, Barra, and Silué 2011). Even though Guei, Barra, and Silué (2011), did not study the crop-raiding element, their principle was that early maturity could be a vehicle for mitigating stress. In Mozambique, short cycle crop varieties were planted to facilitate early harvesting (Rusinamhodzi et al. 2012), thus reducing the risk of losing crops to pests.

Although the studies mentioned earlier have identified the relationship between crop raiding and climatic factors, especially rainfall, a major component of crop raiding that has not been critically examined in the quest to reduce crop raiding in Botswana is manipulating the crops themselves. Very little attention was paid to aligning cropping duration to high rainfall, thus minimizing incidences of crop raiding. Similarly, late supply of seeds, aggravating the misalignment of rainfed planting time with rainfall is a major concern in the eastern Okavango. Affected communities and governments use lots of resources to counter-medicate the actual crop raiding rather than address factors influencing the onset of crop raiding. These counter mitigations do not bode well with the country’s mission toward achieving SDG 1: no poverty and SDG 2: zero hunger (United Nation Botswana 2021). Many researchers (Graham and Ochieng 2008; Hoare 1999; Jackson et al. 2008; Osborn and Parker 2003) have strongly suggested that dealing with consequences or symptomatic treatments of Human-Elephant Conflict (HEC) has not been very productive or successful. These previous reports, therefore, called for a preventative rather than symptomatic treatment of HEC.

This study capitalized on the principle which intended to manipulate crop maturity to minimize the incidence of crop raiding. Unfortunately, breeding newer varieties could be costly and time-consuming, so commonly available varieties were used. Maize, sorghum and beans are major crops in the Ngamiland East (Statistics Botswana 2017), and often have varieties that mature quickly. In Botswana these early varieties were bred for drought escape and short rainfall durations. Therefore, this study presumes that early maturing varieties escaping drought could also evade elephant raids by reducing the exposure time crop stay in fields. Successful escape of early maturing cultivars from pests may encourage easy adoption since such crops are part of the indigenous foods for farmers in the region.

This study proposed the avoidance or prevention strategy via early harvest rather than the symptomatic treatment of elephant crop raiding. The objective of this study is to evaluate how the maturity cycle of crops influences the farmers’ vulnerability to elephant crop raiding. The questions raised by the study were (i) could planting early maturing varieties minimize the likelihood of crop raiding? (ii) does variation in crop maturing timelines influence the risk of crop raiding differently? (iii) could crop maturing timelines be used as a mitigation measure to alleviate crop raiding? and (iv) could early maturing varieties improve food availability amongst subsistent farmers? This evaluation could reduce farmers’ vulnerability to poverty by saving them the food losses incurred due to elephant crop raiding. Counter mitigation processes can be both time and resource consuming, therefore the study could lower the costs associated with counter controlling elephant raiding.

Materials and methods

The comparison experiments were conducted in the Ngamiland District of Botswana in 2018. The study was centered in four selected villages (Figure 1) on the Okavango Delta Panhandle, namely Eretsha (18° 46` 25.77``S, 22° 42’08.74’‘E), Gunotsoga (18° 49` 29.71``S, 22° 35’28.23“E), Mogotho (18° 30` 51.83``S, 22° 08’40.99“E) and Sekondomboro (18° 21` 03.26 “S, 21° 57’38.287“E). Rainfall in this area reaches around 120 mm to 210 mm per year (Botswana Department of Meteorological Surveys 2016). Mean minimum and maximum temperatures are 25°C and 35°C, respectively (Statistics Botswana 2015).

Figure 1. Geographical location of the study sites indicating the four villages; Sekondomboro, Mogotho, Eretsha and Gunotsoga in the Okavango Delta eastern Panhandle, where the studies were conducted.

The Okavango Delta is protected internationally and locally through bodies such as the IUCN; World Heritage Site, and a Ramsar site (International Union for Conservation of Nature 2010). Locally, the NG11 and NG12 are designated Wildlife Management Areas and Controlled Hunting Areas (CHAs) in the Ngamiland district. As a result, there is a wide variety of megaherbivore species in the Okavango. However, elephants are more abundant and have been steadily increasing in numbers, posing a major threat to the woodlands around the Delta (Ramberg et al. 2006). Since 1994, Botswana’s elephant population has been increasing, and their range has expanded by 43% (Songhurst 2010). Given the expansion of range occupied and utilized by elephants, there is an increased likelihood of HEC occurring around human populated areas. Communities in these villages include ethnically diverse people, mainly from Basarwa, Bayei and Hambukushu tribes. They practise small-scale, rain-fed arable and subsistence farming as a primary source of livelihood sustenance (Motsholapheko, Kgathi, and Vanderpost 2011). Apart from subsistence agriculture, the majority of people work in the tourism related industry, such as lodges and tourist satellite villages (Ramberg et al. 2006) or practise fishing and crafting artifacts. Farmers practice rain fed arable farming and rely heavily on Government subsidy schemes for their farming inputs (Motsumi, Magole, and Kgathi 2012). Therefore, their arable agriculture is often characterized by low productivity and instability, mainly arising from arid climate and poor soils (Totolo 1999).

Experimental design and treatments

On-farm experiments were set up concurrently in the four villages for one cropping season; from October to April 2018, to assess early and late maturing varieties of sorghum, maize and cowpea to elephant exposure. Farmers in the eastern Okavango primarily rely on Government for seeds, hence, readily available varieties of these crops were sought from the Department of Agricultural Research. The early maturing varieties were sorghum PSL985028 (59–65 days), maize ZM309 (90–120 days) and cowpeas ER7 (60 days). For comparative purposes, late maturing varieties used were Segaolane maturing at around 120 days, CZH0623 maturing at 130 days, and Tswana cowpeas maturing at 110 days.

The eastern Okavango area is a resource constrained area, and conducting land demanding experimental trials or acquiring virgin land for trials can prove very difficult. Many farmers are increasingly abandoning their farmlands due to elephant crop raiding. This study, therefore, mutually worked with farmers on their fields on a Farmer-Researcher managed basis. The setup allowed for the researcher to propose and provide test crops, whilst farmers implemented the cultivation as per their usual management practises. The implementation included planting until harvest and the researcher mainly collected data during the monitoring. The Farmer-Researcher managed setup allows for an increased buy-in from farmers as they get to appreciate the difference brought about by the mitigation. Moreover, allowing farmers to use their management practices improves mitigation adoption rates and sustenance possibilities in the future.

This study adopted and modified the experimental design in Gross et al. (2016), who proposed alternative crops to alleviate elephant crop raiding. However, in our study, a “multi location trail,” randomized complete block design (RCBD) was explored. The RCBD experimental plot measuring 546 m² (Figure 2), constituted of three replications within it. Each replication had randomly allocated test crops or treatments appearing once in it. The experimental plot was then repeated in four different locations in the eastern Okavango, hence multi location RCBD. The experimental plot replications consisted of experimental units or treatments measuring 4 m × 6 m (24 m²), with a spacing of 0.5 m between treatments. Early and late varieties of maize, sorghum and cowpeas were allocated randomly within the experimental units. The inter-row and intra-row spacing for sorghum, maize and cowpea were 0.75 m x 0.3 m, 0.9 m x 0.5 m and 0.75 m x 0.2 m as guided by the field crops reference handbook of Botswana (Manthe-Tsuaneng and Maphanyane 2000). The experiment was conducted under rain-fed conditions, and varieties were sown on the same day in December. However, the varieties were harvested as and when they reached maturity. In order to conform to the majority of local farmers’ practices, the trial did not use any fertilizers, herbicides or pesticides, and weeding was done using a hand hoe. During the cropping season, the, fields were monitored every day until all the crops reached maturity and were harvested. The researcher was responsible for the elephant interaction assessment and recordings.

Figure 2. Part of the experimental plot layouts showing some of the crops at the gunotsoga (left) and mogotho (right) locations. The early and late maturing varieties of maize, cowpea and sorghum were placed in a RCBD setup.

Data collection and analysis

Yields for crops were determined by harvesting, shelling, and then weighing. Thereafter, the yield was converted from kg to tonnes per hectare (t ha⁻¹). The evidence of elephants in and around the experimental plots was recorded daily, and elephant markers observed within 100 m of the farms were also recorded, even if the elephants did not enter the experimental plots. Markers used to attest to the presence included spoors such as elephant prints and dung, elephant sightings and damaged crops as prescribed in the IUCN data collection protocol Hoare (1999). The elephants’ presence data was later aggregated into weekly recordings for easier management and the total number of days elephants were around the crop during their entire stay in the field. The elephant presence was converted into percentages (elephant sightings %) and then correlated with the duration of crops in the field. Therefore, crop exposure and vulnerability to raiding were associated with the presence of elephants whilst the crop was still standing in the field. There was a need to correlate rainfall to crop raiding and crop maturity timelines and decide on the ideal threshold for farmers to evade crop raiding and obtain higher yields. Accumulated field raids assessment data from the Ecoexist office for the period 2008 to 2018 was used to estimate long-term crop raiding. The rainfall and temperature data for the 2008 – 2018 period were acquired from the Meteorological Department.

ANOVA and General Linear Models (LM) were used to determine any variance in the findings (yield) and a Post Hoc comparison of means was done to compare yield from early and late maturing varieties of the same crop to identify whether the results between different crops or locations are similar or different. Regression analysis at a 95% confidence interval was used to communicate the associations and influence of crops stay in the farm and the vulnerability or exposure to elephants. In this case, crop stay or duration was an independent variable and exposure to elephants was the dependent variable. Moreover, Spearman’s correlation analysis was also computed to show the relationship between temperature and rainfall (independent variables) to crop raids (dependent variable) at 0.05 significance level. using, General Linear Models (LM) and. Thereafter, a maturity category presenting a lower risk and exposure to elephant raiding, hence better evasion to raiding and yielded better was recommended for use by farmers. The computations were done using the Statistical Analysis System (SAS) and the statistical package for IBM Social Sciences (SPSS) version 23.0.0.

Results

Changes in elephant presence with time

Elephant presence significantly increased with the duration of crop stay in the fields (Figures 3). Reducing crop stay in the field potentially reduced crop susceptibility and exposure to elephant raiding. Late maturing varieties presented higher exposure to elephants, whilst early maturing cultivars were less exposed therefore less susceptible to crop-raiding risk by elephants. The exposure risk was three times higher for late maturing varieties compared to the early maturing counterparts.

Figure 3. Elephant presence around the experimental plot during the cropping season, calculated at a 95% confidence interval. The image denotes the time when the crop was harvested. At all locations, the intensification of elephants increased as the cropping season advanced and as crops reached maturity. Early maturing cowpea (EC) and sorghum (ES) were harvested at around week 10 and 11. Early maturing maize (EM) was harvested at around weeks 15–18. late maturing maize (LM), sorghum (LS) and cowpea (LC) were harvested at aroundield variation between early and late maturing cultivars at different locations.

By the time of harvesting late-maturing varieties, elephants were recorded as present around the field every day of the week (100% sightings). In contrast, elephants were only recorded for 40%, 57.1% and 57.14% of the time for early maize, sorghum and cowpea respectively.

Yield variation between early and late maturing cultivars at different locations

In general, the yield from the four sites differed significantly (p < 0.001). The Duncans grouping indicates that higher yields were observed in Sekondomboro (0.48 t ha⁻¹) and Gunotsoga (0.38 t ha⁻¹), whereas Mogotho and Eretsha presented lower yields at 0.14 t ha⁻¹ and 0.21 t ha⁻¹ respectively. Early maturing varieties yielded more than their late counterparts (Figure 4) except for maize grown at Sekondomboro and Mogotho. The yield values are reported in Figure 4 and Appendix 1. Compared with cowpea and maize, sorghum generally produced lower yields. When tested at p < 0.05, the statistical evidence suggests that maturity duration did not significantly impact yield (p > 0.05). Similarly, the interaction between crop type and maturity (p > 0.05) then maturity and location (p > 0.05) did not influence yield. However, the interaction of crop type, location, and maturity duration significantly impacted yield (p < 0.01).

Figure 4. Comparison of yield between early and late maturing varieties of sorghum, maize and cowpea. Early varieties of sorghum and cowpea yielded higher; 0.2 t ha⁻¹, compared to 0.05 t ha⁻¹ for sorghum, 0.31 t ha⁻¹ compared to 0.17 t ha⁻¹ for cowpeas. Early maturing maize yielded 0.53 t ha-¹ and 0.55 t ha⁻¹ for the late-maturing cultivar.

There was an apparent variation in yield between early and late varieties for cowpeas (p < 0.05) and sorghum (p < 0.05). The mean yield for early maturing cowpeas was 0.31 t ha⁻¹ compared to 0.17 t ha⁻¹ for late maturing cowpeas. Early maturing sorghum produced higher yields (0.2 t ha⁻¹) compared to late maturing sorghum (0.05 t ha⁻¹). In contrast to cowpea and sorghum, there was no statistical difference (p > 0.05) between early-maturing (0.53 t ha-¹⁾ and late-maturing maize cultivars (0.55 t ha⁻¹).

Cumulative influence of rainfall, temperature and maturity timelines on crop raiding within a cropping season

Despite the high susceptibility of the area to elephant crop raiding, elephants did not raid the on-farm trials. However, they were mainly seen on the peripheries of the farm, moving to and from the Delta channels. Cumulative crop raiding data from Ecoexist and climatic data from the meteorological department were used to explain the general influence of rainfall and temperature on elephant raiding during the cropping period (Figure 5).

Figure 5. Long term monthly variation in rainfall, temperature and elephant raiding during the cropping season (october-june) from 2008 to 2018. Elephant crop raids are significantly influenced by and dependent on rainfall as the season advances. However, the temperature effect is not as pronounced as that of rainfall. Crop raiding intensifies as the rainfall subsides and during the onset of the dry season.

As rainfall increased, less crop raiding was recorded; however, crop raiding significantly increased as the rainfall amounts reduced (Figure 5). There was a negative correlation between rainfall and crop raiding (r = 0.35, n = 125 p = 0.00), whereas the influence of temperature on raiding was not significant (r = 0.02, n = 125, p = 0.81). These results suggest that rainfall is a primary regulator of crop raiding rather than temperature. The 2017/2018 rainy season was high after two consecutive low rainy seasons. This led to increased rains coinciding with this on-farm trial, thus accounting for the elephant presence but limited elephant raid incidents.

From 2008–2018, there was nearly no crop raiding in October (0), November (1) and December (0), and this period correlated with the seedling stage, and the early vegetative stages of the crops. The peak crop raids were observed in February (331), March (368), April (403) and May (139). Few raids were recorded in June (9). This period translates to a period within the two dotted vertical lines (Figure 5) when rainfall starts decreasing and many crops are at the late vegetative, flowering and physiological maturity stages. Many early maturing cultivars ripen by the first cross (just by or after February), while many late maturing ripen around the second cross (by or around April) when raiding was at its peak. Rainfall amounts presented an increasing trend from an average of 10.3 mm, 48.0 mm, 74.1 mm, and 158.9 mm for October, November, December, and January, respectively (Figure 5). However, it then declines from 96.0 mm, 88.1 mm, 42.6 mm, 3.2 mm, and 4.9 mm in February, March, April, May and June months, respectively. The crop raiding records show a lagging effect on the rainfall as depicted in the Figure 5.

Discussion

Association between the number of raids, climatologic factors and crop maturity timelines

The evidence in Figure 5 shows a consistent increase in the number of raids as the cropping season advanced. The presentation in the figure correlates with the strong statistical significance levels depicted in Figure 3 for the trial locations. This result aligns with Masunga (2007) who reported that elephant movement and crop raids highly correlate with water availability. When rainfall amounts increased, the data show that little to no raids were recorded, especially between October and December when many farmers plant or the crops would be at the seedling or early vegetative stages. Similarly, June had fewer crop raids, as many farmers would have already harvested, therefore there were no crops to depredate. Jackson et al. (2008) and Parker and Osborn (2001) had previously reported a similar pattern. None of the farm trials were raided during the 2017/2018 cropping season. Statistics Botswana (2017) had forewarned and even reported that the season had above average rainfall compared to the 2015/16 season. This high rainfall phenomenon is consistent with our results in the 2017/2018 period and reinforces the fact that elephant raids subside in the season of high rainfall. Moreover, most of the 2017/2018 rains fell late in January and mostly in February. This rainfall, however, reduced toward the maturity of many late maturing cultivars and was followed by consecutive high temperatures.

The long-term climatic data results suggest that when rainfall amounts decrease, crop raiding significantly increases. Under most instances, as the rainfall reduces, temporary water sources in the bush dry up, making the elephants move to access drinking water from the permanent sources such as the Okavango Delta channels (Masunga 2007). Most of the elephants access farms on their way to and from the water sources (Parker and Osborn 2001). Songhurst (2012) also reported that elephants tend to travel long distances during drought or pre-rainfall season in search of water. As they move, there is an increased likelihood of invading crops. By those times, crops would be at the most desirable stage for elephants, as they would have started fruiting, with some almost attaining physiological maturity. Furthermore, Warner (2008) has articulated that peak crop raiding occurs since the natural forage would have started drying out and being more fibrous; therefore, elephants would pursue fresh, supple crop supplies. Parker, Osborn, and Hoarse (2007) have attributed this peak raiding to fruiting bodies of crops being the most nutritious compared to natural forage at the latter stages of the rain season. This peak corresponds to February, March, April and May in Figure 5, and corresponds with weeks 8–20 in Figures 3. The above findings suggest that farmers should target harvesting by February and or shortly after February. Leaving crops any further than mid-February worsens the farmer’s risk of losing the crops. In all four locations, early maturing varieties were harvested before the increased presence of elephants near the fields. However, late maturing varieties were exposed to an increased risk of crop raiding due to the prolonged stay in the fields, which coincided with high incidences of elephant presence. Consequently, the timelines translate to increased vulnerability of farmers to crop loss because of increased crop stay in the fields. Thus, a safer period for many farmers to secure their harvest was mid-February (Figure 5), when many early maturing cultivars were ready for harvest. Even though many subsistence farmers in the Okavango are inclined toward traditional crop varieties due to reasons, such as ease of access and the ability to reuse the seed in the future, the traditional varieties mature late at around 120 days (April-May). Hence, the increase in crop raids with longer crop maturity duration emanating from the onset of the dry season coupled with the ripening of crops and the intensification of elephants’ presence around fields. Owing to this combination, farmers’ crops remain a much better and more nutritious alternative for elephants during the dry season period. Hence, many farmers end up in heightened poverty and food insecurity (Mayberry, Hovorka, and Evans 2017).

Comparison of crop yields

The yield of each crop varied from one location to another. Similarly, the interaction of crop, location and maturity timeline had a significant impact on yield. Apart from climatological conditions, this association highlights the effects of other probable factors, such as soil fertility variations in these different locations, as they can influence the yield of crops. Charles et al. (2010) also reported that factors such as varying climatic conditions and soil properties at the different locations affect the growth and maturity rate of many crops in rain-fed agriculture, resulting in yield gaps. Jackson et al. (2008) emphasized similar sentiments regarding rainfall or temperature. However, the experimental locations in this study are not too far from each other, and the replications and randomization were done to counter such confounding factors where available.

An apparent variation where early maturing varieties yielded more than the late maturing varieties could be attributed to the evasion of pests and harsh environmental factors. The degree of exposure to possible damage by elephants and non-elephant factors in the field is possibly attributed to the differences in yield between the two maturity groups during the trials. Guei, Barra, and Silué (2011) have reported that crops, which mature early incur less exposure and damage from pests and unfavorable environmental factors. The results align with Egbe, Alibo, and Nwueze (2010), where early maturing cowpeas had significantly higher yields, especially in cereal intercropping. The use of hybrid varieties which mature early has been outlined by (Charles et al. 2010) as another means of improving yields in challenging environments. Even though the challenging environment in their study referred to climatic challenges, the purpose of earlier harvesting also fits very well in areas where elephants are a major challenge. Using appropriate hybrid varieties to curb climatic impacts was further reported by Guei, Barra, and Silué (2011), as crops maturing early, and not harvested late were less likely to be affected by drought and pests. This timing consequently led to increased yield and food availability compared to crops that stayed longer on the farm.

Evading environmental stresses such as extreme heat and pests, especially around reproductive development could be the probable cause for early cultivars yielding better than late cultivars in this study. Abundant nutrition and moisture are very important for crop growth and yield (Tittonell and Giller 2013). Even though temperature plays a role in agricultural production, rainfall has proven to be Africa’s most dominant climatic element influencing yield (Müller et al. 2011). Similarly, the impact of rainfall has been very evident in this study. Our findings concur with Wossen et al. (2017) and Oyinbo et al. (2019) who reported increased yields in early varieties of maize mainly due to a reduction in vulnerability to drought risk. For cowpea and sorghum, which matured by February, adequate rainfall was guaranteed to support reproductive development, hence the higher yield potential for these cultivars. Egli (2011) also discouraged planting of cultivars with longer total growth duration as they devote much of their resources to vegetative growth, consequently decreasing the yield. In this study, the early and late maturing maize cultivars used, had only ten days difference in maturity times. Thus, the early cultivar reached physiological maturity at around the same time as its late counterpart. This slight variation in maize maturity days implied that both early and late cultivars incurred almost similar vulnerability to environmental stresses and other non-elephant pests. Consequently, the response showed by the early maize cultivar is an artifact of cultivar choice. This study strongly affirms that the availability of a much earlier maturing cultivar could have produced similar results to cowpea and sorghum.

Conclusion

The inundation of the eastern Okavango panhandle by large numbers of elephants poses a severe impediment to subsistence food production and elephant conservation. A single day of crop exposure to elephants could mean complete crop loss to elephants. The results suggest that planting crops that mature before the onset of the dry season and elephant peak raids reduces the time and extent of crop elephant interactions. This crop maturity manipulation effectively increases the potential to secure the harvest, therefore reducing poverty among smallholder farmers in high human elephant conflict areas. On a similar note, late maturing crop varieties significantly increased exposure of crops to elephant raiding, and late maturing varieties yielded lower than their early maturing counterparts. Therefore, the findings presented substantial evidence that early maturing crop varieties could be strategically considered and included as an effective mitigation to improve the cropping system resilience in the eastern Okavango. The study is optimistic that further climate action advancements in this trans-disciplinary approach could yield more results toward attaining zero hunger and no poverty among smallholder farmers. Unlike already existing measures, this noninvasive and affordable mitigation measure will ensure the evasion of elephants and reduce the farmers’ vulnerability to food loss. To sustain traditional varieties, the early varieties could be grown alongside the traditional varieties to help cushion farmers in cases of late maturing crop loss due to elephant raiding. Collectively, the ability of farmers to securely harvest crops with minimal impact on biodiversity fosters positive agroecological practices, peaceful co-existence and reduces the retaliatory killing of elephants by farmers.

Acknowledgments

This work was supported by the Ecoexist Project and Botswana University of Agriculture and Natural Resources. We are grateful to the farmers who granted us access to conduct assessments in their fields and participated in the monitoring of fields.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data are stored at Botswana University of Agriculture and Natural Resources library https://library.buan.ac.bw, Gaborone, Botswana and Ecoexist Trust offices in Maun. A public deposit is available at Matsika, Tiroyaone Albertinah (2022), “Elephant & crop maturity comparisons”, Mendeley Data, V1, doi: 10.17632/mxhvhjfbxc.1

Additional information

Funding

The work was supported by the Ecoexist Project.

Appendix 1.

Yield data analysis

The SAS System 10: 36 Wednesday, October 10, 2019 10

The GLM Procedure

Table

Crop	Maturity type	Location	N	yield
				Mean	Std Dev
Cowpea	Early	Eretsha	3	0.28	0.21
Cowpea	Early	Gunotsoga	3	0.43	0.35
Cowpea	Early	Mogotho	3	0.27	0.13
Cowpea	Early	Sekondomboro	3	0.26	0.21
Cowpea	Late	Eretsha	3	0.10	0.05
Cowpea	Late	Gunotsoga	3	0.29	0.14
Cowpea	Late	Mogotho	3	0.10	0.04
Cowpea	Late	Sekondomboro	3	0.17	0.08
Maize	Early	Eretsha	3	0.41	0.25
Maize	Early	Gunotsoga	3	0.84	0.70
Maize	Early	Mogotho	3	0.07	0.07
Maize	Early	Sekondomboro	3	0.81	0.17
Maize	Late	Eretsha	3	0.33	0.15
Maize	Late	Gunotsoga	3	0.41	0.15
Maize	Late	Mogotho	3	0.3	0.43
Maize	Late	Sekondomboro	3	1.17	0.33
Sorghum	Early	Eretsha	3	0.10	0.05
Sorghum	Early	Gunotsoga	3	0.17	0.14
Sorghum	Early	Mogotho	3	0.09	0.14
Sorghum	Early	Sekondomboro	3	0.42	0.19
Sorghum	Late	Eretsha	3	0.057	0.03
Sorghum	Late	Gunotsoga	3	0.11	0.02
Sorghum	Late	Mogotho	3	0.00	0.00
Sorghum	Late	Sekondomboro	3	0.04	0.04

Table

Class Level Information
Class	Levels	Values
rep	3	1 2 3
Location	4	1 2 3 4
Crop	3	1 2 3
maturity	2	1 2

Table

Number of Observations Read	72
Number of Observations Used	72

The SAS System 10: 36 Wednesday, October 10, 2018 11

The GLM Procedure

Dependent Variable: Yield (tonha)

Table

Sum of
Source	DF	Squares	Mean Square	F Value	Pr > F
Model	25	5.79760139	0.23190406	4.65	<.0001
Error	46	2.29513056	0.04989414
Corrected Total	71	8.09273194

Table

R-Square	Coeff Var	Root MSE	Yieldtonha Mean
0.716396	74.07940	0.223370	0.301528

Table

Source	DF	Type I SS	Mean Square	F Value	Pr > F
rep	2	0.21633611	0.10816806	2.17	0.1260
Location	3	1.27217083	0.42405694	8.50	0.0001
maturity	1	0.13956806	0.13956806	2.80	0.1012
Crop	2	2.22750278	1.11375139	22.32	<.0001
Locatio*Crop*maturit	17	1.94202361	0.11423668	2.29	0.0133

Table

Source	DF	Type III SS	Mean Square	F Value	Pr > F
rep	2	0.21633611	0.10816806	2.17	0.1260
Location	3	1.27217083	0.42405694	8.50	0.0001
maturity	1	0.13956806	0.13956806	2.80	0.1012
Crop	2	2.22750278	1.11375139	22.32	<.0001
Locatio*Crop*maturit	17	1.94202361	0.11423668	2.29	0.0133

Appendix 2.

Post hoc comparison of yield early and maturing crop yeilds

Table

Number of Means	2	3	4
Critical Range	.1533	.1612	.1664

Table

Means with the same letter are not significantly different.
Duncan Grouping	Mean	N	location
A	0.47778	18	Sekondomboro
A	0.37556	18	Gunotsoga
B	0.21389	18	Eretsha
B	0.13889	18	Mogotho

Table

Means with the same letter are not significantly different.
Duncan Grouping	Mean	N	Crop
A	0.54125	24	Maize
B	0.23917	24	Cowpea
B	0.12417	24	Sorghum

Table

Crop	Level of Maturity	N	yield Mean	Std Dev
Cowpea	Early	12	0.31250000	0.21418450
Cowpea	Late	12	0.16583333	0.10974833
Maize	Early	12	0.52916667	0.46421896
Maize	Late	12	0.55333333	0.44898539
Sorghum	Early	12	0.19500000	0.18128330
Sorghum	Late	12	0.05333333	0.04830459

Appendix 3.

Regression analysis on elephant sightings as cropping time advances

Table

The REG Procedure
Dependent Variable: Sekondomboro sightings
Number of Observations Read	16
Number of Observations Used	16

Table

Analysis of Variance
Sum of Mean
Source	DF	Squares	Square	F Value	Pr > F
Model	1	14616	14616	240.39	<.0001
Error	14	851.20903	60.80065
Corrected Total	15	15467

Table

Root MSE	7.79748	R-Square	0.9450
Dependent Mean	48.19938	Adj R-Sq	0.9410
Coeff Var	16.17755

Table

Parameter Estimates
Parameter Standard
Variable	DF	Estimate	Error	t Value	Pr > |t|
Intercept	1	−7.53125	4.08903	−1.84	0.0868
time	1	6.55654	0.42288	15.50	<.0001

**********************************************************.

Table

The REG Procedure
Dependent Variable: Eretsha sightings
Number of Observations Read	16
Number of Observations Used	16

Table

Analysis of Variance
Sum of Mean
Source	DF	Squares	Square	F Value	Pr > F
Model	1	10937	10937	46.46	<.0001
Error	14	3295.64002	235.40286
Corrected Total	15	14233

Table

Root MSE	15.34284	R-Square	0.7685
Dependent Mean	51.78563	Adj R-Sq	0.7519
Coeff Var	29.62761

Table

Parameter Estimates
Parameter Standard
Variable	DF	Estimate	Error	t Value	Pr > |t|
Intercept	1	3.57550	8.04586	0.44	0.6636
time	1	5.67178	0.83208	6.82	<.0001

**********************************************************

Table

Dependent Variable: Gunotsoga
Number of Observations Read	16
Number of Observations Used	16

Table

Analysis of Variance
Sum of Mean
Source	DF	Squares	Square	F Value	Pr > F
Model	1	11190	11190	100.63	<.0001
Error	14	1556.74475	111.19605
Corrected Total	15	12746

Table

Root MSE	10.54495	R-Square	0.8779
Dependent Mean	54.45688	Adj R-Sq	0.8691
Coeff Var	19.36386

Table

Parameter Estimates
Parameter Standard
Variable	DF	Estimate	Error	t Value	Pr > |t|
Intercept	1	5.69450	5.52982	1.03	0.3206
time	1	5.73675	0.57188	10.03	<.0001

**********************************************************

Table

Dependent Variable: Mogotho sightings
Number of Observations Read	16
Number of Observations Used	16

Table

Analysis of Variance
Sum of Mean
Source	DF	Squares	Square	F Value	Pr > F
Model	1	10469	10469	106.47	<.0001
Error	14	1376.62090	98.33006
Corrected Total	15	11845

Table

Root MSE	9.91615	R-Square	0.8838
Dependent Mean	49.98438	Adj R-Sq	0.8755
Coeff Var	19.83850

Table

Parameter Estimates
Parameter Standard
Variable	DF	Estimate	Error	t Value	Pr > |t|
Intercept	1	2.81850	5.20007	0.54	0.5963
time	1	5.54893	0.53778	10.32	<.0001

Appendix 4.

Table

Correlations
			Rainfall_mm	Crop_raids	Temperature_oC
Spearman’s rho	Rainfall_mm	Correlation Coefficient	1.000	−.345**	.548**
		Sig. (2-tailed)	.	.000	.000
		N	125	125	125
	Crop_raids	Correlation Coefficient	−.345**	1.000	.022
		Sig. (2-tailed)	.000	.	.806
		N	125	125	125
	Temperature_oC	Correlation Coefficient	.548**	.022	1.000
		Sig. (2-tailed)	.000	.806	.
		N	125	125	125

** Correlation is significant at the 0.01 level (2-tailed).