Urban Pollution Reshapes Insect Life Along a Semi-Arid River in Botswana

A small river winding through Botswana’s capital city is offering scientists an unusually clear picture of how urbanization rewrites the rules of freshwater life. The Segoditshane River, which drains much of Gaborone before joining the Notwane River, has become the focus of a new study showing that the chemical and physical fingerprints of a growing city act as powerful ecological filters, selectively removing sensitive aquatic insects while allowing hardy, pollution-tolerant species to persist. The research, published in the journal Environmental Challenges, is among the first to apply an environmental filtering framework to a semi-arid urban river in southern Africa, a region where most freshwater research has concentrated on celebrated systems such as the Okavango Delta rather than on the smaller, heavily stressed waterways that run through expanding cities.

The study was conducted by Thomamo B. Benjamin, Kelemogile Mmolawa, Gaolathe Tsheboeng, Marks Ditlhogo, and Francis O. Arimoro, who sampled three fixed stations along the river between February and May 2026. The stations were deliberately chosen to represent a gradient of urban influence: an upstream reach near the A1 Highway with relatively low-density development, a midstream reach cutting through the most densely built-up part of the city, and a downstream reach near the Tlokweng Road corridor that integrates the cumulative drainage of the entire urban catchment. Because the Segoditshane flows through residential, commercial, institutional, and industrial zones, it functions as the city’s principal stormwater conduit, collecting runoff, wastewater influence, and diffuse pollutants along its roughly 14-kilometer corridor through the metropolitan area.

The physicochemical results reveal a river under measurable strain. Dissolved oxygen fell from a mean of 7.59 milligrams per liter at the upstream station to roughly 4.8 to 5.0 milligrams per liter at the midstream and downstream sites, a decline of about 36 percent that the authors link to rising organic loading. Electrical conductivity climbed from around 734 to 742 microsiemens per centimeter upstream and midstream to 952 microsiemens per centimeter downstream, an increase of nearly 30 percent that reflects the accumulation of dissolved salts, nutrients, metals, and other urban-derived ions. Biochemical oxygen demand rose from 3.28 milligrams per liter upstream to 8.75 midstream, chemical oxygen demand more than doubled from 18.5 to 45 milligrams per liter, and total dissolved solids increased from 455 to 590 milligrams per liter downstream. Nitrate concentrations also rose steadily along the continuum, while phosphate peaked at the midstream station.

These chemical shifts tracked a semi-quantitative urban pressure assessment developed by the team. The upstream site scored 3 on a ten-point pressure scale, with roughly 25 percent impervious surface cover, a surrounding population density of about 1,500 people per square kilometer, and only one or two major stormwater outfalls. The midstream site, embedded in the densest urban fabric, scored 8, with approximately 70 percent impervious cover, 4,500 people per square kilometer, and five to seven outfalls. The downstream site scored highest at 9, combining about 60 percent impervious cover, seven to ten stormwater outfalls, high wastewater influence, and full integration of upstream drainage. The close agreement between these pressure scores and the measured water chemistry suggests that catchment development and drainage connectivity are central to the environmental degradation documented in the river.

Against this backdrop, the biological story is striking. Over four monthly sampling campaigns, the team collected 1,238 macroinvertebrate individuals representing 16 families from five insect orders, identified to family level using South African Scoring System Version 5 taxonomic keys. Abundance collapsed along the urban gradient, from 718 individuals upstream to 436 midstream and just 84 downstream. The upstream station was dominated by pollution-tolerant Diptera, particularly non-biting midges of the family Chironomidae with 250 individuals and mosquitoes of the family Culicidae with 156, alongside moderate numbers of predatory water bugs and damselflies. The midstream station supported a more balanced assemblage including mayflies of the family Baetidae, dragonflies and damselflies, and several families found nowhere else in the study, such as Gomphidae, Micronectidae, Scirtidae, and Gyrinidae. The downstream reach held only six families, with Baetidae, Chironomidae, Libellulidae, and a handful of others accounting for nearly all individuals.

Diversity metrics added an important nuance to this pattern. Although the upstream station recorded the highest raw abundance, its Shannon diversity of 1.844 was the lowest of the three sites, a consequence of dominance by a few opportunistic taxa. The midstream station, by contrast, posted the highest values for nearly every metric: 11 families, a Margalef richness index of 1.645, Shannon diversity of 2.265, Simpson diversity of 0.886, and Pielou’s evenness of 0.945. This midstream diversity maximum is one of the study’s most notable findings, suggesting a non-linear response along the river continuum. Rather than a simple slide from good to bad conditions, the pattern points to intermediate disturbance dynamics, in which moderate environmental stress and greater habitat heterogeneity at the midstream reach widen the range of niches available, allowing sensitive and tolerant taxa to coexist while competitive exclusion is dampened.

A modified application of the South African Scoring System reinforced this interpretation. The midstream station achieved a SASS score of 58, twelve scoring taxa, and an average score per taxon of 4.83, corresponding to a condition of Good under the SASS5 framework. The upstream station scored 35 with an ASPT of 3.89, rated Fair, while the downstream station fell to a score of 23 and an ASPT of 3.83, rated Fair to Poor. The authors caution that because their pooled multi-habitat sampling differed from the standardized SASS5 kick-sampling protocol, these indices should be read as comparative indicators of relative condition rather than formal ecological status classifications. Even so, the convergence of abundance, richness, diversity, and biotic index data on the same spatial pattern lends weight to the overall conclusion.

To connect community structure with environmental conditions, the researchers employed Canonical Correspondence Analysis, a multivariate ordination technique that relates species distributions to measured gradients. Based on twelve station-by-month observations, the analysis showed that the first two canonical axes together explained 97.82 percent of the constrained family-environment variation, with Axis 1 accounting for 65.72 percent and Axis 2 for 32.10 percent, both statistically significant under Monte Carlo permutation tests. Axis 1 loaded most strongly on ammonia, pH, dissolved oxygen, depth, and phosphate, while Axis 2 was associated with electrical conductivity, flow velocity, depth, nitrate, and phosphate. Because several environmental variables were strongly correlated with one another, the authors interpret the ordination as an exploratory description of interacting gradients rather than a test of independent drivers. Upstream assemblages aligned with higher pH and oxygen, midstream communities with flow, depth, and nutrients, and downstream assemblages with elevated conductivity and nitrate.

The findings fit squarely within the theory of environmental filtering, which holds that local abiotic conditions act as a sieve, favoring species whose traits allow survival under prevailing conditions while excluding those that cannot cope. In the Segoditshane River, the filters include oxygen depletion, nutrient enrichment, rising salinity indicated by conductivity, altered flow and depth, and habitat simplification driven by impervious surfaces and channel modification. Tolerant families such as Chironomidae and Culicidae thrived where organic enrichment and standing water created favorable breeding conditions, while moderately sensitive predators like Gomphidae and Aeshnidae were largely confined to the midstream reach, and the relatively sensitive Baetidae and Simuliidae persisted mainly where localized hydraulic refugia offered patches of suitable habitat. Importantly, the authors warn that the higher proportional contribution of sensitive taxa downstream should not be mistaken for recovery; it more likely reflects the loss of other groups from an already depauperate community.

The study carries practical weight for a rapidly urbanizing semi-arid region where conventional chemical monitoring alone can underestimate cumulative ecological degradation. Because macroinvertebrates integrate environmental stress over weeks and months, they provide a biological memory that instantaneous water samples lack. The authors argue that their results establish a baseline for biomonitoring in the Segoditshane River and support integrated catchment management spanning water-quality protection, stormwater management, habitat restoration, and long-term ecological monitoring. They also outline future priorities, including year-round monitoring across the full hydrological cycle, greater spatial replication, habitat-specific sampling, trait-based and species-level approaches, and the incorporation of environmental DNA metabarcoding to sharpen taxonomic resolution. As Gaborone continues to grow, the Segoditshane’s insect communities may prove to be among the most honest witnesses to the city’s environmental trajectory.

Subject of Research: Urban pollution signatures and environmental filtering of benthic macroinvertebrates along abiotic gradients in a semi-arid river in Botswana

Article Title: Urban pollution signatures and environmental filtering of benthic macroinvertebrates along abiotic gradients in a semi-arid river in botswana

Article References: Benjamin, T. B., Mmolawa, K., Tsheboeng, G., Ditlhogo, M., & Arimoro, F. O. (2026). Urban pollution signatures and environmental filtering of benthic macroinvertebrates along abiotic gradients in a semi-arid river in botswana. Environmental Challenges, 25, Article 101661. https://doi.org/10.1016/j.envc.2026.101661

Image Credits: AI Generated

DOI: 10.1016/j.envc.2026.101661

Keywords: benthic macroinvertebrates, environmental filtering, urban pollution, semi-arid river, water quality, biomonitoring, SASS5, dissolved oxygen, electrical conductivity, Botswana, Gaborone, canonical correspondence analysis