A Hippo in the Room: Predicting the Persistence and Dispersion of an Invasive Mega-vertebrate in Colombia

Background / Context

  • Study focused on the invasive population of the common hippopotamus (Hippopotamus amphibius) in Colombia’s Magdalena River basin.
  • Hippos are native to sub-Saharan Africa (29 countries, 115,000–130,000115{,}000\text{–}130{,}000 individuals) but were illegally introduced in the 1980s by drug trafficker Pablo Escobar (1 male + 3 females).
  • Species currently listed as Vulnerable by IUCN; Colombian population is escalating and dubbed the world’s largest invasive animal.
  • Conflict: charismatic appeal vs. ecological, social, and economic damages; no formal inclusion under Colombia’s National Invasive Species Act.

Key Terminology & Concepts

  • Invasive Alien Species (IAS): species introduced by humans outside native range, able to establish self-sustaining populations and cause socio-environmental impacts.
  • Population Viability Analysis (PVA): stochastic simulation to forecast population persistence/extinction under varying demographic & management parameters.
  • Ecological Niche Modelling (ENM): statistical prediction of suitable habitat based on species–environment relationships; used here with ensemble algorithms and future climate scenarios.
  • Carrying Capacity (KK): upper limit of individuals an environment can sustain; estimated at K=1500K = 1500 hippos for the Magdalena system.
  • Representative Concentration Pathways (RCP): greenhouse-gas concentration trajectories used for climate projections. RCP 2.6 = low-emission, RCP 8.5 = high-emission.

Study Area – Magdalena River Basin

  • Largest Colombian watershed (≈257,438 km2257{,}438\,\text{km}^2; ≈25 % of national territory).
  • Length ≈1600 km1600\,\text{km}, headwaters at ∼3800 m\sim3800\,\text{m} a.s.l.; >80 % of Colombian population lives within basin.
  • Seasonal flood pulse creates a mosaic of rivers, tributaries, marshes (locally “ciénagas”).
  • Floodplain municipalities house ≈3.35 million3.35\text{ million} inhabitants; ~50 % have Unsatisfied Basic Needs.

Methods – Population Viability Analysis (PVA)

  • Software: VORTEX v10.3.8; 500 iterations per scenario.
  • Founder group: 1 ♂ + 3 ♀ released 1993.
  • Demographic inputs (selected examples):
    • Sex ratio at birth 1:11{:}1
    • Age at 1st offspring: ♀ 5 y5\,\text{y}, ♂ 6 y6\,\text{y}
    • Max reproductive age: ♀ 43 y43\,\text{y}, ♂ 45 y45\,\text{y}
    • Offspring per brood: 1 (96 %) or 2 (4 %); max 1 brood yr⁻¹
    • Adult female breeding probability: 80 %
    • Mortality (captive-based): calf ♀ 28 %, calf ♂ 31 % (0–1 yr); adults ≥3 yr 1 %
    • Mate monopolization: 80 % mature ♂
  • Management scenarios (500 simulations each | 100-yr horizon):
    1. Baseline – no action.
    2. Current sterilization rate (0.23 ind yr⁻¹).
    3. Increased sterilization: 6, 10, 16 ind yr⁻¹ (Scen 2–4).
    4. Extraction (culling/translocation): 10, 20, 30 ind yr⁻¹ (Scen 5–7).

Methods – Habitat Suitability / ENM

  • Occurrence data: confirmed hippo localities (white dots on Fig. 1); outlier Magangué sighting excluded.
  • Environmental predictors: 7 non-collinear WorldClim bioclimatic variables (Bio2, 3, 4, 6, 15, 18, 19).
  • Modelling ensemble: MaxEnt, MARS, GBM, RF, CART, SVM, GLM, GAM.
  • Validation: 70 % train / 30 % test, pseudo-absences (train = 10× presences, test = 100× presences), metrics = TSS + omission rate.
  • Future transfer: 16 GCMs × 2 RCPs (2.6 & 8.5) × 2 time slices (2050, 2070).
  • Dynamic dispersal simulation (cellular automata):
    • Two dispersal kernels: flat Pdisp=1P_{disp}=1 vs. negative exponential.
    • Propagule production: constant vs. age-increasing (maturity from 5 yr to full maturity 43 yr).
    • 95-year time series (2005–2100) example shown with CCSM4 RCP 8.5.

Results – PVA

  • Back-calculated 2020 population size: 98±4398 \pm 43 hippos (≈4 new hippos yr⁻¹ since 1993); extinction probability 0.2%0.2\%.
  • Baseline (no control):
    • Growth rate r=0.145r = 0.145 (≈69 new hippos yr⁻¹).
    • Half KK (≈783±59783 \pm 59) reached 2034; full KK (≈1418±1441418 \pm 144) by 2039.
  • Sterilization scenarios:
    • 0.2 ind yr⁻¹: identical to baseline → full KK 2039.
    • 6 ind yr⁻¹ (r ≈ 0.135) → full KK 2043.
    • 10 ind yr⁻¹ (r ≈ 0.131) → full KK 2045.
    • 16 ind yr⁻¹ (r ≈ 0.115) → full KK 2053; minor eradication probability (0.2 %).
  • Extraction scenarios:
    • 10 ind yr⁻¹: r ≈ 0.132; full KK 2044.
    • 20 ind yr⁻¹: initial decline to 38±3538 \pm 35 hippos by 2042, rebound to full KK 2074.
    • 30 ind yr⁻¹: negative growth r=−0.218r = -0.218; eradication predicted 2033.

Results – Habitat Suitability & Climate Change

  • Ensemble projections: vast suitable habitat <$1500\,\text{m}$ across northern Colombia.
  • Range size expands almost linearly to 2100 under full dispersal.
  • RCP 2.6: slight expansion by 2070.
  • RCP 8.5: major expansion; suitability extends to Sierra Nevada de Santa Marta (north) & Tolima lowlands (south).
  • Dynamic models concur: without dispersal limits, entire Magdalena-Cauca basin becomes occupiable before 2100.

Discussion – Population Growth & Expansion

  • Estimated annual growth 14.5 % aligns with high-growth African sites (0–18 % variation).
  • Optimal Colombian conditions: abundant forage, permanent water, no predators, minimal human pressure.
  • Early sexual maturity noted (♂ sperm active at 4 yr), accelerating growth.
  • Previous field counts (≤80 hippos) likely underestimated due to cryptic behavior and methodological gaps.

Ecological Implications

  • Nutrient loading: 50 % of hippo feces/urine directly into water → higher N\text{N} & P\text{P} → eutrophication, cyanobacterial blooms.
  • Physical disturbance: wallowing & sediment agitation alter geomorphology, hydrology; may hinder migratory fauna (e.g. Antillean manatee).
  • Disease reservoir: anthrax, brucellosis, Rift Valley fever, trypanosomiasis, schistosomiasis, etc.

Economic & Social Implications

  • Livelihood threats: fisheries decline (water quality + space competition), crop/livestock damage, infrastructure destruction.
  • Human-wildlife conflict: hippos cause more human fatalities in Africa than any other large mammal; first serious Colombian attack recorded 2020.
  • Indirect costs: guarding effort, loss of income, psychosocial stress, food security risks.

Potential Benefits? – ‘Empty Niche’ Debate

  • Argument: hippos fill ecological roles of extinct Pleistocene toxodonts.
  • Counterpoints:
    • Toxodont semiaquatic habit unproven; isotope data suggest terrestrial.
    • Native semiaquatic herbivores already present (Antillean manatee, capybara).
    • Founder effect (1 ♂ + 3 ♀) → low genetic diversity; unsuitable for species conservation/re-wilding in Africa.

Management Strategies Evaluated

  • Sterilization/Castration
    • Low rates ineffective; ≥30 ind yr⁻¹ (≥50 % females) needed merely to delay growth.
    • Costly (surgery, anesthesia, logistics), risky, low cost-benefit.
  • Extraction (capture or culling)
    • Capture & translocation feasible only near Hacienda Nápoles; requires secure enclosures (electric fences, trenches).
    • Culling (≥30 hippos yr⁻¹ starting 2021) is sole scenario achieving eradication (by 2033).
    • Social resistance strong; hippos viewed as tourist attraction in Doradal.

Monitoring & Research Recommendations

  • Implement systematic aerial dry-season surveys; supplement with drones, eDNA, bioacoustics, camera traps.
  • Establish participatory local monitoring networks in middle & lower Magdalena.
  • Conduct studies on genetic diversity, health, and pathogen load.
  • Quantify socio-economic costs of hippo presence through long-term field and modelling approaches.
  • Improve public engagement: communicate ecological risks vs. charisma; incorporate stakeholders in decision-making.

Conclusions & Policy Implications

  • Without intervention, population will surpass 14001400 hippos by 2039, with basin-wide colonization enhanced by climate change.
  • Current low-level sterilization ineffective; only high-level extraction/culling can halt and reverse invasion.
  • Delay in action heightens ecological degradation, human conflict, and economic loss.
  • Urgent need to revise Colombian regulations to permit effective control, backed by robust scientific evidence and transparent public outreach.