Data from: Perceived and actual ecosystem services by fruit bats, birds, and primates in litchi orchards agroecosystems
Data files
Aug 25, 2025 version files 21.47 KB
-
Mphethe-JAPPL-2025.xlsx
15.54 KB
-
README.md
5.93 KB
Abstract
Ecosystem services and disservices by key animal groups such as birds, fruit bats, and primates are often understudied. The lack of ecological knowledge results in persecution and culling due to perceived damage to fruit trees. We combined a control-treatment study with questionnaire data to untangle the actual and perceived impact of birds, fruit bats, and primates in South African litchi orchards. Control trees produced significantly higher yields than the caged trees. There was a weak, but non-significant, trend towards higher vertebrate damage on control trees compared to treatment trees. Questionnaire data found that weather and monkeys caused more damage to litchi fruits compared to fruit bats. There was also strong consensus among farmers that damage varied between vertebrate (monkey, fruit bats, birds, and wild pigs) and invertebrate (insects) as well as an abiotic factor (weather) groups. Farm type (commercial versus small-scale) showed a statistically significant difference in perceived damage caused by vertebrate groups and weather. Commercial farms reported higher damage caused by groups when compared to small-scale farms. This study recorded relatively low incidences of crop raiding by fruit bats and highlights the economic benefits of biocontrol by insectivorous bats and birds, which outweighed the yield losses by fruit bats, birds, and monkeys. Vertebrate exclusion to prevent crop damage limits access and biocontrol benefits provided by bats and birds. Farmers are unaware of the economic benefits of bats and birds; thus, it is vital to educate them on the ecological importance which can outweigh the disservices resulting from fruit bats and reduce persecution by fruit farmers.
Dataset DOI: 10.5061/dryad.41ns1rnsz
Description of the data and file structure
This dataset includes experimental exclusion and questionnaire survey data collected to investigate the impacts of vertebrate exclusion on litchi agroecosystems and to assess farmer perceptions of litchi crop damage. An exclusion experiment was conducted over two consecutive harvest seasons (2019–2020) on two litchi farms in South Africa. Six nylon mesh cages (2×3 cm mesh size) were used to exclude flying vertebrates (e.g., bats and birds), monkeys, and large herbivores, while allowing access to arthropods and small animals. Two treatments were applied, a Full exclusion (closed at all times) and Control (open trees with no cage). Each exclusion cage was paired with an open control resulting to three paired treatments per farm. Cages were installed three months before the 2019 harvest and removed after the 2020 season.
In parallel, a structured survey was conducted with 51 litchi farmers (13 commercial and 38 small-scale) to assess their perceptions of litchi crop damage caused by vertebrates, invertebrates, and abiotic factors. The survey collected data on farm characteristics, yield losses, fruit bat activity, management practices, economic impacts, and attitudes toward bats and other wildlife. Ethical approval was obtained and all participants provided informed consent.
Files and variables
File: Mphethe-JAPPL-2025.xlsx
Description: Contains data from the vertebrate exclusion experiment conducted on two litchi farms during the 2019 and 2020 harvest seasons. It also includes responses from 51 litchi farmers (13 commercial, 38 small-scale) on perceived crop damage, management practices, and bat-related attitudes.
Variables (Exclusion experiments data)
| Variable Name | Description |
|---|---|
| Farm | Unique identifier for each farm (e.g., F1) |
|---|---|
| Tree Name | Unique identifier for each tree in the study |
| Treatment | Exclusion treatment applied: FULL (cage), CONTROL (open) |
| Year | Year of observation (2019 or 2020) |
| First grade (kg) | High quality fruits that meet the top standards for commercial sale and export |
| Second grade (kg) | Fruits that do not meet top standards but considered commercially acceptable for sale |
| Rejects (kg) | Fruit that failed to meet commercial or consumption standards due to spoilage and are discarded |
| Total yield | Total amount of fruit harvested from the caged and control trees |
Variables (Questionnaire data)
| Variable Name | Description |
|---|---|
| Location | Area where the farm is located |
| Farmer Type | Commercial or Small-Scale |
| Litchi Yield (kg) | Estimated annual litchi yield (kg) |
| Fruit Bat Raids | Frequency of fruit bat activity reported (e.g., every night) |
| Bat Damage Percent | Estimated % of yield lost to bats |
| Other Damage Groups | Other sources of crop damage (e.g., Birds, Monkeys) |
| Bat Management Methods | Practices used (e.g., netting, noise deterrents) |
| Litchi Skin | Number of lichi skin found on the ground during litchi season (e.g., plentiful) |
| Consulted Expert | Whether the farmer consulted an expert on bat damage (Yes, No) |
| Attitude Bats | Farmer's general attitude (e.g., Positive, Neutral, or Negative) |
| Bat Ecosystem Services | Whether the farmer is aware bats provide ecological benefits |
| Preferred Control Method | Farmer's view on most effective control method |
| Bat perception | Farmers perception towards bats (good/bad) |
Code/software
All data analyses and visualizations were conducted using R, a free and open-source statistical computing environment. The data can be viewed and processed using R. although the analyses and results presented in this study were performed using R version 4.4.1 (R Development Core Team, 2024).
Required Software:
- R (version 4.4.1 or later) from https://www.r-project.org
- RStudio (optional but recommended for ease of use)
Main R Packages Used:
- lme4 – For fitting linear and generalized linear mixed-effects models
- lmerTest – For obtaining p-values and t-tests for mixed-effects models
- glmmTMB – For fitting Tweedie GLMMs and other advanced models
- DHARMa – For model diagnostics and residual checking
- MuMIn – For model selection and AICc calculation
- ggplot2 – For data visualization
- dplyr, tidyr – For data wrangling
- car – For additional statistical tools
Human subjects data
All participants gave informed consent and were assured of anonymity and confidentiality throughout the study
