Data from: Are crop fields pharmacies for megaherbivores? From ecophysiological studies of elephant (Loxodonta cycotis) crop raiders in Gabon

Ngama, Steeve 1 2 ; Bindelle, Jerome2; Brown, Janine3; Vermeulen, Cédric2

Research facility: Smithsonian Conservation Biology Institute

Published Sep 19, 2025 on Dryad. https://doi.org/10.5061/dryad.7wm37pw1p

Data files

Sep 19, 2025 version files 22.93 KB

Forest_elephant_crop_raiding_according_to_FGCM_and_GIP.csv

17.74 KB
README.md

5.19 KB

Abstract

Damage to crops is a major cause of human-elephant conflict (HEC) in elephant range states. Elephant crop raiding drives farmers' resentment against elephants and reduces local community support for wildlife conservation. While elephant crop raiding ecology is well studied, further investigations on HEC mitigation strategies are still needed. Thus, there is a need to focus on less investigated areas, such as the physiological drivers of elephant crop-raiding behavior, using multidisciplinary sciences. Two physiological proxies, gastrointestinal parasite infestations (GPI) and fecal glucocorticoid metabolite (fGCM) concentrations, common in animal ecophysiology, were used to help understand differences or motivations in the preferences of crops by elephant raiders. The results show, for the first time, that forest elephants may increase the frequency of crop raiding according to GPI, indicating a self-medication behavior. Increases in parasitism prevalence (PP) and parasitism intensity (PI) in sampled boluses led to 28% and 0.16% more intakes of all crops, respectively. Parasitism prevalence (PP) increases in elephant boluses also led to 16% and 25% more bananas and papaya intakes, respectively, while PI increases in boluses led to 0.1% more intakes of both bananas and papaya plants. No such predictions were found for other crops (cassava and palm plant), nor for natural food species. Furthermore, fGCM concentrations were not related to elephant crop raiding. Results highlight a trade-off between the benefit of elephants raiding crops and the danger of encountering farmers by adopting nocturnal crop-raiding behaviours.

Dataset DOI: 10.5061/dryad.7wm37pw1p

Description of the data and file structure

Forest elephant crop raiding according to FGCM and GIP

Dataset Overview

This data includes frequencies of forest elephant crop raiding events as well as fecal glucocorticoid metabolite concentrations and gastrointestinal parasitism presence. The data is composed by 90 rows corresponding to individual elephants sampled. Columns represents variables with Fecal Glucocorticoid concentration, parasite prevalence (PP), and parasitism intensity (PI) being the response variables and the others being the explanatory variables.

Funding

These data were collected through a research work funded by COLAS-Gabon with the collaboration of IRAF-CENAREST (Gabon), Nature+ asbl NGO (Belgium), and Gembloux ABT, University of Liège (Belgium). COLAS-Gabon, IRAF-CENAREST, and Nature+ provided logistic and financial support for the field work to SN. Gembloux Agro-Bio Tech, University of Liège (Belgium), and RESSAC (CIFOR-ICRAF) provided financial support for data analyses, manuscript writing, and publishing.

Ethics Approval

The National Center for Scientific Research and Technology (Centre National de la Recherche Scientifique et Technologique, CENAREST) and the National Agency of National Parks (Agence National des Parcs Nationaux ANPN) in Gabon provided permissions to conduct this research.

Files and variables

File: Forest_elephant_crop_raiding_according_to_FGCM_and_GIP.csv

Description: fecal glucocorticoid metabolite concentrations (fGCM), as well as parasite prevalence (PP) and parasite intensity (PI), are the response variables. The others parameters are explanatory variables.

Variables

Seasons: Our observational experimental blocks were seasons with five modalities (S1=Long dry season; S2 = short wet season; S3 = short dry season; S4 = long wet season; S5= second long dry season)
S5: second long dry season
S4: long wet season
S3:short dry season
S2:short wet season
S1: Long dry season
Elephant.Groups:
Crop: represents the presence (1) and absence (0) of farmed plants eaten by elephants.
Crops: represent all farmed plants eaten by elephants.
Food.sp: represents all plant species eaten by elephants
Nat.food: is a numerical form of non-croped plants eaten by elephantsNat.food.: is the numerical form of non-croped plants eaten by elephants
Natural.food: represents all non-crop plants eaten by elephants
Banana: is a crop species eaten by elephants
Banana.: is the numerical form of the banana crop species eaten by elephants
Cassava: is a crop species eaten by elephants
Cassava.: is the numerical form of the cassava crop species eaten by elephants
Papaya: is a crop species eaten by elephants
Papaya.: is the numerical form of the papaya crop species eaten by elephants
Palm: is a crop species eaten by elephants
Palm.: is the numerical form of the palm crop species eaten by elephants
Costus: is a natural species eaten by elephants
Cost.: is a natural species eaten by elephants (numeric)
Bamboo: is a natural species eaten by elephants
Bamboo.: is a natural species eaten by elephants (numeric)
Fern: is a natural species eaten by elephants
Fern.: is a natural species eaten by elephants (numeric)
Ficus: is a natural species eaten by elephants
Ficus.: is a natural species eaten by elephants (numeric)
Parasitism: indicates the presence (1) or absence (0) of parasite eggs in elephant dung
PP: is parasitism prevalence
FGM.conc: fecal glucocorticoid metabolite concentrations
FGMcx10: fecal glucocorticoid metabolite concentrations times 10 for graphical depictions
FGM.conc/SEM: fecal glucocorticoid metabolite concentrations for graphical depiction
EPGD.c: egg per gram of boluses
PI:Parasitism intensity
EPGDx100: egg per gram of boluses times 100 (for graphical representations)
EPG.cat: egg per gram of boluses (categorised)
OPG.cat.num: egg per gram of boluses (numeric)
PP.ord: Parasitism prevalence (ordinal for analyses)
OPG.cat.L.H.num: egg per gram of boluses (categorised for analyses)

Code/software

We performed descriptive and inferential statistical calculations using Excel software and R version 4.0.4 (R Development Core Team 2025). We performed descriptive statistics, including percentage calculations, charts’ design, and principal component analyses (PCA) showing correlations between variables. In inferential statistics, the Shapiro test was used to assess the normality of response variables, while the X² dependency test was used to assess whether variables were dependent. As most of our data were counts, we used nonparametric ANOVA tests (Kruskal and Wilcoxon) to calculate PP, PI, and fGCM differences according to plant intakes. We also used the lme4 package to perform generalized linear mixed models (GLMMs) analyses to predict the effects of PP, PI, and fGCM on elephant plant species intakes over time.