Effects of agroforestry practices on insect abundance, richness, and diversity, in transformed ecosystems of the semi-arid area, Dodoma, Tanzania

1 Introduction

The rapid conversion of forests into agricultural ecosystems, settlements, and infrastructure to meet growing population demand has accelerated biodiversity loss worldwide, diminishing the ecological services provided by biota (Bos et al., 2007; Hausmann et al., 2022). The management of these transformed landscapes will determine whether global biodiversity conservation efforts succeed (Gardner et al., 2009). Conservation strategies must therefore prioritize biodiversity within altered ecosystems, particularly in semi-arid regions where ecological resilience is fragile (Musgrave, 2013). Sustainable agroecological practices that mimic natural ecosystems, such as agroforestry, have been promoted as viable options to enhance species diversity and ecosystem stability (Donald and Evans, 2006).

Agroforestry, the integration of trees into agricultural land, offers multiple environmental, social, and economic benefits (Michael, 2013). By conserving and diversifying tree species on farms, agroforestry helps sustain biodiversity in human-dominated landscapes (Schroth et al., 2004). Evidence suggests that agroforestry systems can support up to 60% of the species diversity found in natural forests (Edo et al., 2024; Schroth et al., 2004). This capacity stems from the structural complexity created by multipurpose trees, shrubs, and grasses, which provide diverse niches and resources for insect communities. Despite this ecological importance, limited research has examined how agroforestry influences insect diversity and abundance in transformed semi-arid ecosystems such as Dodoma, Tanzania.

Insects are key indicators of ecosystem health because they are sensitive to habitat change and play roles in pollination, decomposition, and trophic interactions (Lami et al., 2023; Maleque et al., 2009; Ngongolo and Nyundo, 2015; Nuraeni et al., 2026; Uehara-Prado et al., 2009). Researchers have widely used their diversity and abundance to assess ecological integrity in wetlands, rangelands, and forests (Mouhoubi et al., 2019). Semi-arid landscapes, however, present unique constraints: low rainfall, high temperatures, and sparse vegetation that shape insect community dynamics. In such environments, agroforestry may play a disproportionately important role by enhancing habitat heterogeneity, buffering microclimatic extremes, and providing continuous floral and structural resources that sustain diverse insect guilds (Perfecto and Vandermeer, 2010; Sollai and Solari, 2022; Villamarin-Cortez et al., 2026).

Although many studies have documented insect diversity across natural forests, wetlands, and monocultures, fewer have focused on the conservation potential of newly established agroforestry systems in semi-arid regions. This gap is critical because agroforestry is increasingly promoted as a climate-resilient land-use option in drylands (Hausmann et al., 2022; Lindenmayer and Laurance, 2017). Understanding its role in sustaining insect communities is therefore essential for designing land-use systems that maintain ecological functioning and resilience. This study addresses this gap by examining how agroforestry practices influence insect abundance, richness, and diversity in transformed ecosystems within the University of Dodoma. By comparing insect communities across agroforestry plots, afforested sites, shrubland, and bushland, we evaluate the conservation value of agroforestry relative to other semi-arid habitats and highlight its potential contribution to biodiversity management.

2 Materials and methods

2.1 Description of the Study Area

The research was conducted at the University of Dodoma (UDOM), located in Ng’ong’ona ward, Dodoma district, Dodoma Region (Figure 1). The University is about 8 km east of the city and covers about 6,000 acres. UDOM is situated in a semi-arid area with low, unimodal annual rainfall of approximately 550–600 mm between December and April (Gayo, 2022). With an annual average temperature of 29.0 °C, the mean daily temperature is lowest in July (13.1 °C) and highest in November (30.6 °C) (Kayombo et al., 2020). Dry savanna shrub-thicket vegetation with sporadic trees and grassland areas characterizes semi-arid areas of Tanzania. During the dry season, the bush becomes leafless, while during the wet season it flourishes with abundant life, making the whole area fully green (Namwata et al., 2015). The study sites were selected from the four colleges of the University of Dodoma (UDOM) as described in Table 1. The three selected sites have had their original ecosystems transformed into alternative land uses after infrastructure construction, making them suitable for studying insect community structure compared with the fourth site, which is an intact, undisturbed semi-arid forest ecosystem. Gayo (2022) described the native and exotic vegetation. The dominant native vegetation is Bussea massaiensis, Acacia spp, Terminalia sericea, Brachystegia spiciformis, Sclerocarya birrea, Commiphora spp, Maerua decumbens, Grewia forbesii, and Euphorbia candelabrum. The common exotic species include Leucaena leucocephala, Tamarindus indica, Mangifera indica, Afzelia spp, Asarinas spp., Eucalyptus spp, Trichilia emetica, Peltophorum pterocarpum, Khaya anthoteca, Cedrela audorata, Zyzigium cuminii, and Combretum apiculatum,

Table 1. Description of the study sites.

2.2 Insect sampling

Insect sampling was conducted in March/April 2025, when vegetation is fully leafy and green, and most species are in their flowering or seed-production stage. During this period, a remarkable amount of litter also accumulated, allowing different surface-active insect activity. In semi-arid areas, surface-active insect activity peaks during this period, just before the dry season; therefore, the literature recommends sampling during this transitional period (Davies et al., 2015). Insects were collected from four study sites: areas transformed by establishing trees alongside crops (here referred to as agroforestry), areas transformed by replacing natural ecosystems with mostly exotic trees (here referred to as afforested), intact bushy areas (here referred to as bushland), and areas transformed by clearing the bush but with few remaining shrubs with or without grasses (shrubland). Each site was divided into two transects, each 100 m long, spaced 500 m apart (Gayo, 2022; Mwambala and Nyundo, 2016; Stanbrook et al., 2021). We placed five 1-L plastic pitfall traps along each transect at 20-meter intervals, for a total of 10 pitfall traps per sampling site. We set a total of 40 traps across the study area. The pitfall traps were filled with 250 ml of water containing a few drops of detergent to break the surface tension, and no bait was used, as we needed to collect different insect species (Ngongolo and Nyundo, 2015). To maximize trapping effort (sample size), we visited each trap regularly, as suggested by Mwambala and Nyundo (2016). Over three consecutive days, we emptied pitfalls every 24 hours and reset them for the next day, yielding 120 samples over 2880 trapping hours. The insects were temporarily stored in a plastic collecting bag, preserved in 70% alcohol (Marquina et al., 2021), and then transported to the College of Natural and Mathematical Sciences (CNMS) laboratory for sorting and identification. We identified collected insects to species level using standard taxonomic keys and regional guides (Martins, 2015), supplemented by a South African guidebook (Picker et al., 2004) and mobile applications such as Google Lens and iNaturalist. We grouped unidentified specimens into morphospecies based on distinct morphological traits. We included these morphospecies in diversity and ordination analyses. An experienced expert from the Department of Biology, University of Dodoma, guided the procedures. Preserved insect samples form part of a teaching specimens’ collection in the CNMS Biology laboratory.

2.4 Data analysis

Descriptive statistics including frequencies and percentages were used to summarize surface-active insect species abundance and richness in the study area. We assessed site diversity using the Shannon-Wiener and Simpson indices. Simpson’s index can be biased when more abundant species are added to a sample, and the Shannon-Wiener index tends to give greater weight to the addition of a new species to a sample, even if it is rare. Therefore, both Simpson’s and Shannon-Wiener indices were used to counterbalance the limitations of each other in describing the diversity of habitats sampled, as suggested by Stanbrook et al. (2021). We used generalized linear models (GLMs) to assess habitat effects on insects’ abundance, richness, and diversity indices. We performed non-metric multidimensional scaling (NMDS) with the function metaMDS() in the vegan package 2.7-2 (Oksanen et al., 2025), using Bray–Curtis dissimilarities to visualize differences in community composition among habitats (Stanbrook et al., 2021). We tested compositional differences with PERMANOVA using adonis2(), and assessed within-community variability with betadisper() followed by permutest(). All permutation tests used 999 permutations, with no restrictions imposed. Significance was determined at P < 0.05. All analyses were conducted in R version 4.5.2 (R Core Team, 2024).

3 Results

3.1 Insect Abundance

Based on 120 trap nights, a total of 2,889 insects were collected from the study area (Appendix 1). Of these, the most abundant insects were found at afforested sites (30%, n=863), followed by agroforestry sites, shrubland, and bushland (Table 2). In the study area, Aranea spp were most common (29.70%, n=858), followed by Pachycondyla tarsata (27.13%, n=784) and Lepisiota spp. (26.83%, n=775). Amitermes spp., Epioscopomantis chalybea, Gryllacnididae spp., Ceroctis spp., Japygidae spp., Stenocara spp., Meliponula spp., and Mymicaria natalensis were the least numerous species in the research areas, with n=2, 0.04%, each. Based on the relative abundance of each species at a given location, Pachycondyla tarsata (17.44%) and Lepisiota spp. (12.32%) were found to be particularly numerous. Both bushland and shrubland had higher abundances of Araneae spp. (7.68%) and (10.18%), respectively.

Table 2. Summary of insect abundance, species richness, and diversity across habitats.

3.2 Species richness

Across the research area, we recorded 53 insect species from 14 orders (Appendix 1). The shrubland had the highest number of species, followed by the afforested land area, while the bushland and agroforestry each had the lowest (Table 1). Lepisiota spp., Pachycondyla tarsata, and Araneae spp. are among the species with high distribution in the research area (Appendix 1). Amitermes spp., Epioscopomantis chalybea, and Gryllacrididae spp., by contrast, are less common in the research area.

3.3 Insect species diversity

Among the four habitats assessed, shrubland exhibited the highest Shannon diversity (1.86), followed by bushland, which showed similarly elevated Shannon diversity and the highest Simpson index after shrubland, reflecting a relatively low-dominance (higher evenness) structure. In contrast, Forested sites supported intermediate diversity, while agroforestry systems had the lowest diversity across all indices (Table 1).

3.3.1 Effects of habitat types on insects’ abundance, richness and diversity

The generalized linear model results (Table 3) showed that habitat type measurably affected surface-active insect abundance, richness, and diversity across the study sites. The intercept estimates confirmed that afforested habitats had a significantly higher baseline abundance, richness and diversity. By contrast, none of the alternative habitats showed a significant positive effect (Table 2).

Table 3. Effects of habitat type on insect abundance, richness, and diversity.

Note: Estimates are shown relative to the afforested reference, with standard errors, z-values, and p-values<0.05.

3.4 Insect community composition

The non-metric multidimensional scaling (NMDS) ordination (Figure 2) revealed clear separation of surface-active insect communities across the studied habitats. The stress value (0.068) indicated an excellent fit of the two-dimensional solution, meaning the 2-dimensional NMDS solution captures the community dissimilarities very reliably. To formally test differences in community composition, we conducted a PERMANOVA using Bray–Curtis dissimilarities. The analysis showed that habitat explained 41.8% of the variance in community composition, but the effect was not statistically significant (R² = 0.42, F = 0.96, p = 0.546), indicating that compositional differences among habitats were weak and not supported by permutation testing.

In contrast, the test for homogeneity of multivariate dispersions (PERMDISP) revealed highly significant differences in dispersion among habitats (ANOVA: F ≈ 1.34 × 10³¹, p < 0.001; permutation test: p < 0.001). This result shows that although habitats did not differ significantly in centroid location (community composition per se), they varied strongly in sample spread within groups. In other words, insect assemblages exhibited heterogeneous variability across habitats, suggesting that differences in community structure are driven more by dispersion than by distinct compositional shifts.

4 Discussion

4.1 Species abundance

The higher abundance of surface-active insects observed in areas afforested with trees may be attributed to the resource-rich microhabitats created in the woodland plots. Unlike agroforestry sites, woodland areas maintain lower vegetation with minimal disturbance, providing conducive environments for insect communities to survive. Moreover, replacing bushy vegetation with properly arranged trees in woodland plots may have destroyed higher-predator habitats, offering insects peaceful living areas and thereby enhancing their abundance. This observation aligns with Kimambo et al. (2020) and Msanga et al. (2018), whose results showed that afforestation enhanced insects abundance through the provision of microhabitats within the afforested areas. Gayo (2022) found that afforested sites at the University of Dodoma were ideal for many beetle species, consistent with the current study, which assessed broader insect communities in these areas. However, the finding that agroforestry and shrubland supported a declining abundance agrees with those of Olarewaju et al. (2022), which claimed that converting natural ecosystems to alternative land uses leads to declines in insect populations. Further literature indicates that variations in environmental conditions, the time of year insects were collected, and the presence of vulnerable hosts at the research location can all be linked to changes in observed abundance (Shakeel et al., 2019). The age of the transformed ecosystem, though not assessed in this study, affects the system’s ability to support insect abundance and diversity (Chan et al., 2008; Gayo, 2022), supporting this study’s observation that agroforestry contributes to declining insect abundance. The high abundance observed in afforested woodlands underscores their conservation value as resource-rich refuges for insect populations in semi-arid landscapes. Protecting and expanding these woodlands, while minimizing disturbance, should therefore be prioritized to sustain insect abundance and ensure the continuity of ecosystem services.

4.2 Species richness

Regarding species richness, although shrubland had the highest species counts, regression analysis confirmed that only afforested woodland habitats contributed significantly to species richness, whereas shrubland showed a positive but non-significant effect, and agroforestry and bushland had no measurable influence. The apparent richness in shrubland may therefore reflect local sampling variation or transient habitat conditions rather than a consistent ecological effect. By contrast, the structural complexity of afforested sites appears to provide stable microhabitats that sustain diverse insect assemblages. These results diverge from Nagy et al. (2016) and Medina and Lopes (2014), who reported higher insect richness in shrublands and bushlands due to complex food web structures. Similarly, Gayo (2022) observed greater richness in shrublands driven by the coexistence of omnivorous, phytophagous, and predatory beetles. This discrepancy may reflect differences in habitat age, successional stage, and the presence of residual natural vegetation. In our study, woodland plots provided resources such as leaf litter, shade, and microclimatic buffering, which expanded available microhabitats and mimicked natural forest conditions, supporting higher richness than alternative land-use types. From a conservation perspective, the higher richness in structurally complex afforested plots underscores the importance of maintaining tree cover and microhabitat diversity. These habitats provide stable refuges for insects and should be prioritized in semi-arid restoration strategies.

4.3 Species diversity

The study found that introducing habitats that differed greatly from existing ones increased surface-active insect abundance but decreased species diversity. Bushland and shrubland had higher evenness than agroforestry and afforested areas. This pattern can be attributed to non-native vegetation, coupled with more frequent disturbances in agroforestry and afforested sites than in shrublands and bushlands. As a result, a few tolerant species such as Pachycondyla tarsata and Lepisiota spp. appear to dominate the newly established habitats (Appendix 1). This finding is consistent with a study by Gayo (2022), which found higher coleopteran evenness in grasslands and lower in a three-year tree plantation. In addition, Tind Nielsen (2007) reported findings similar to those of the current study; many species of Scarabaeidae and Scarabaeinae preferred wooded patches, such as bushland, over grassland and cultivated sites in semi-arid Tanzania. Similarly, Alarape et al. (2015) reported that animal and insect species diversity corresponds to habitat structural complexity and the diversity of vegetation forms, which aligns with our findings that newly introduced trees affect insect distribution in agroforestry and afforested sites. Interestingly, regression analyses showed that only the afforested site had a significant positive influence on insect diversity indices, meaning that although areas transformed into woodlots harbor a small total number of species, these species are distributed across many insect taxa. This may be because newly established afforested sites include a few predators, such as small mammals. The high diversity in shrubland and bushland underscores their role as reservoirs of insect biodiversity in semi-arid landscapes. Afforested sites, despite intermediate diversity, showed significant positive effects, highlighting their potential as managed habitats that can complement natural systems. By contrast, agroforestry requires targeted interventions, such as integrating native tree species and reducing disturbance, to shift from abundance-driven dominance to balanced diversity. Together, these findings emphasize that conserving natural shrubland and bushland, while improving the management of afforested and agroforestry systems, is critical for sustaining resilient insect communities and ecosystem services.

4.4 Insects’ community composition

The ordination analysis shows partial isolation of insect communities across habitat types. However, those differences have not been sufficient to demonstrate any shifts in community composition. Thus, insect community structure in transformed semi-arid areas appears similar, possibly because comparable resource availability and disturbance factors hinder divergence in community composition. In turn, dispersion analysis has revealed clear differences in within-habitat variability. These results show that habitat heterogeneity, rather than distinct habitat categories, more strongly influences the structure of surface-active insect communities in the studied habitats. This variability underscores the importance of micro-scale habitat features and vegetation patchiness, which can create diverse niches and influence insect diversity patterns (Barton et al., 2009; Tscharntke et al., 2002). Comparable studies in semi-arid agroforestry systems have shown that insect diversity is often driven by structural complexity and resource distribution within habitats rather than by habitat type itself (Munyuli, 2013). Local vegetation structure, canopy cover, and management intensity, though not studied in the present study, have been identified as key factors influencing insect assemblages (Kingazi et al., 2024). In addition, environmental variables such as rainfall, air and soil characteristics, humidity, and litter depth are known to shape insect abundance and community composition, and their absence from the current dataset limits the explanatory power of the ordination. The weak separation but strong within-habitat variability therefore indicates that conservation efforts should focus on enhancing micro-scale heterogeneity and vegetation patchiness. These fine-scale features, together with climatic and soil conditions, are more decisive for sustaining insect communities than broad habitat categories, making them critical management targets in semi-arid landscapes.

4.5 Limitations and future directions

Our study provides important insights into surface-active insect diversity, but several limitations remain. The short sampling period restricted observations to a single seasonal peak, limiting detection of broader temporal dynamics. The exclusive use of pitfall traps, while appropriate for surface-active and crawling taxa, excluded flying insects, which are major pollinators. In addition, vegetation structure and phenological status at each site were not described in sufficient detail, and key environmental variables such as rainfall, air and soil temperature, humidity, canopy cover, litter depth, and soil characteristics were not systematically measured. These omissions constrain interpretation of how habitat conditions may have influenced insect abundance and community composition. Future research should therefore extend sampling across seasons, employ complementary collection methods to capture a wider range of insect guilds, and incorporate detailed vegetation and phenological data. Integrating measurements of climatic and microhabitat factors; including rainfall, temperature, humidity, canopy cover, litter depth, and soil properties would strengthen ecological inference and provide a more comprehensive understanding of insect community dynamics. Such approaches will enhance the explanatory power of agroforestry and ecosystem studies, ultimately offering stronger evidence for biodiversity conservation and management.

5 Conclusion

Agroforestry practices in this study showed a high surface-active insect abundance but reduced diversity, richness, and evenness, with community structure dominated by Pachycondyla tarsata. These findings indicate that while agroforestry enhances surface-active insects’ persistence in semi-arid areas, it may simultaneously constrain ecological balance by favoring dominance over diversity. Results further suggest that habitat type alone exerts limited influence on assemblage separation, whereas within-habitat variation plays a decisive role in shaping surface-active insects’ community composition. These findings underscore that semi-arid agroforestry practices can support surface-active insect survival but require careful management to sustain resilient and functionally diverse communities.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. AI tools like Copilot and ChatGPT were used to write code, interpret data, write the result section, and perform language editing. The draft was then fully rewritten without AI.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fcosc.2026.1918561/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Luis P. Da Silva, Centro de Investigacao em Biodiversidade e Recursos Geneticos (CIBIO-InBIO), Portugal

Reviewed by: Simone Bergonzoli, Council for Agricultural Research and Agricultural Economy Analysis | CREA, Italy

Ngawo Namukonde, The Copperbelt University, Zambia

Francis Ssenkuba, Mbarara University of Science and Technology, Uganda