Climate warming drives vole population collapse by amplifying flea infestation and its physiological costs
Data files
Nov 06, 2026 version files 6.76 GB
-
Figure2_TableS1-S8_raw_data.xlsx
527.14 KB
-
Figure3_raw_data.xlsx
16.31 KB
-
Figure4_raw_data.zip
563.46 MB
-
Figure5_raw_data.zip
6.20 GB
-
Figure6_raw_data.xlsx
11.13 KB
-
README.md
8.72 KB
Abstract
Climate warming is restructuring species interactions, yet its cascading effects through host-parasite systems to entire ecosystems are poorly understood. To predict demographic outcomes and identify responsible mechanisms, we need to establish how warming affects individual physiology and population traits, and how these scale to vegetation dynamics. Integrated, long-term studies dissecting this climate-parasite-host-vegetation pathway are therefore critical but rare.
Here, we test how climate warming affects the interaction between Brandt’s voles (Lasiopodomys brandtii) and their flea parasites, and its consequences for grassland vegetation. We relate a decade of field monitoring data on vole and flea populations to climatic variables, and combine semi-natural enclosure experiments, individual-level controlled laboratory infestation trials, and transcriptomic profiling to isolate causal mechanisms under both field and laboratory conditions.
We show that warmer temperatures significantly intensified flea infestation on voles (ecosystem scale, long-term field monitoring). This increased parasitism led to severe demographic suppression of vole populations, characterized by reduced survival and reproduction (ecosystem and population scale). This parasite-mediated suppression of the dominant herbivore indirectly enhanced grassland vegetation productivity (ecosystem scale, long-term field monitoring). Furthermore, controlled warming synergistically exacerbated the detrimental effects of flea infestation on host survival and reproduction (population scale, semi-natural enclosure experiments).
Results suggest that flea infestation impairs vole fitness through load-dependent physiological costs, including reduced growth and food intake, altered gut microbiota, and disrupted energy allocation and immune-related gene expression (e.g., in protein turnover and IgA production pathways) (individual level, controlled laboratory trials). This establishes that climate warming amplifies the burden of parasitism through interconnected ecological and molecular mechanisms, ultimately restructuring grassland ecosystems (cross-scale synthesis). Our findings highlight a complex climate-parasite-host-vegetation pathway with implications for managing ecosystem health under climate change (cross-scale synthesis).
Dataset DOI: 10.5061/dryad.1rn8pk161
1. Data description
Overview:
This dataset contains the raw data and analysis scripts used to investigate the nonlinear relationships within the climate–parasite–host–vegetation system. Specifically, it explores how climate variables, flea index, rodent density, and NDVI interact, and how flea infestation and warming influence the survival, body mass, immune responses, gut microbiota, and spleen transcriptome of Brandt’s voles (Lasiopodomys brandtii). The dataset includes data underlying Figure 2 through Figure 6, and Supplementary Tables S1-S8.
Data collection and processing:
- Ecological Modeling (Fig 2, Table S1-S8): Generalized additive models (GAMs) and Structural Equation Modeling (SEM) were used to assess nonlinear relationships and causal pathways among climate, parasites, hosts, and vegetation.
- Population and Vegetation (Fig 3): Survival rates, litter size, population size, vegetation coverage, and Leymus chinensis biomass data were collected to evaluate population dynamics.
- Multi-omics and Immunology (Fig 4-5): Serum cytokines (IFN-γ, IL-10, etc.) were quantified using MSD technology. Gut microbiota was analyzed via 16S rRNA sequencing, and spleen transcriptomes were sequenced to assess immune and metabolic responses under flea infestation.
- Survival and Physiology (Fig 6): Kaplan-Meier survival curves, body mass, food intake, and flea survival rates were recorded under different experimental conditions (control, flea-infestation, warmed, and warmed + flea-infestation).
Files and variables
File: Figure2_TableS1-S8_raw_data.xlsx
Description: Raw data and modeling outputs for Figure 2 and Tables S1-S8. This file contains the data used to construct Generalized Additive Models (GAMs) and Structural Equation Models (SEM) to evaluate the nonlinear relationships and causal pathways within the climate–parasite–host–vegetation system.
Variables
- num: Row number; unique identifier for each observation.
- location: study site code: 1 = Abaga (Abag Banner); 2 = East Ujumqin (East Ujimqin Banner); 3 = West Ujumqin (West Ujimqin Banner); 4 = Xilinhot (Xilinhot City), Inner Mongolia, China
- year: Year of sampling (last two digits of the calendar year; e.g. 12 = 2012).
- month: Month of sampling (4 = April; 5 = May; 8 = August; 9 = September)
- z: Flea infestation status of the individual vole (1 = flea-infested; 0 = flea-free)
- tem: Mean monthly air temperature in the sampling month. (C)
- tem1: Mean monthly air temperature with a 1-month lag. (C)
- tem2: Mean monthly air temperature with a 2-month lag. (C)
- pre: Cumulative (total) monthly precipitation in the sampling month. mm
- pre1: Cumulative monthly precipitation with a 1-month lag. mm
- pre2:Cumulative monthly precipitation with a 2-month lag. mm
- rhu: Mean monthly relative humidity in the sampling month. %
- rhu1: Mean monthly relative humidity with a 1-month lag. %
- rhu2: Mean monthly relative humidity with a 2-month lag. %
- ndvi: Normalized Difference Vegetation Index in the sampling month (greenness/vegetation productivity proxy). unitless (−1 to 1)
- ndvi1: Normalized Difference Vegetation Index with a 1-month lag. unitless (−1 to 1)
- ndvi2: Normalized Difference Vegetation Index with a 2-month lag. unitless (−1 to 1)
File: Figure3_raw_data.xlsx
Description: Raw data for Figure 3. This dataset contains population dynamics and vegetation parameters used for statistical comparisons (Kruskal-Wallis and One-way ANOVA) between different experimental groups.
Variables
- first_release_date: Date of the first release of voles. Format: YYYY.M.D or YYYY.M.D-D (e.g., 2024.7.7 or 2024.8.5-6 indicating a date range)
- fenced_plot_id: Identifier for the fenced enclosure plots (C-control group;L-low group;H-high group)
- vole_id: Unique identification number for each individual vole
- sex: Sex of the individual vole; Format: Categorical ("male" or "female")
- body_weight_g: Body weight of the individual vole at the time of recording in grams (A value of 0 likely indicates missing data, unrecorded weight, or individuals not recaptured)
- first_capture_date: Date of the first capture of voles. Format: YYYY.M.D or YYYY.M.D-D (e.g., 2024.7.7 or 2024.8.5-6 indicating a date range)
- second_capture_date: Date of the second capture of voles. Format: YYYY.M.D or YYYY.M.D-D (e.g., 2024.7.7 or 2024.8.5-6 indicating a date range)
Sheet: Field-collected vegetation data:
- Field_enclosure_ID: Identifier for the field enclosures. Format: Categorical (e.g., C1-1, C1-2).
- Stipa capillata, Leymus chinensis:1 = yes
- Total vegetation cover; percentage of total vegetation coverage: Total percentage of vegetation coverage in the plot
- Biomass‑1, above‑ground/below‑ground biomass replicate 1 (gram)
- Biomass‑2, above‑ground/below‑ground biomass replicate 2 (gram)
File: Figure4_raw_data.zip
Description: Raw data for Figure 4. This compressed file contains serum cytokine quantification data (via MSD technology) and fecal 16S rRNA sequencing data (NovaSeq 6000 platform) to analyze immune responses and gut microbiota composition under flea infestation.
File: Figure5_raw_data.zip
Description: Raw data for Figure 5. This compressed file contains spleen transcriptome sequencing data (Illumina NovaSeq X Plus) .
File: Figure6_raw_data.xlsx
Description: Raw data for Figure 6. Contains data for Kaplan-Meier survival analysis and ANOVA comparisons (body mass, food intake, flea survival rates) across four experimental groups: control, flea-infestation, warmed, and warmed plus flea-infestation during the final week of the experiment.
Variables
- Group: Treatment groups (Control, low infestation, high infestation, temperature,temperature + high flea infestation)
- Sex: male and female
- Initial body: Body weight of the individual at the beginning of the experiment
- Body weight (7d/14d/21d/28d): Body weight of the individual measured at day 7, 14, 21, and 28 during the experiment)
All body weight variables (Initial and Days 7, 14, 21, 28) are measured in grams (g).
Note: c = control; l = low infestation;h = high infestation;ct = temperature;ht = temperature + high flea infestation;0 represents death
Code/software
Software used:
- R (Version 3.0.2 and above): Used for GAMs, survival analysis, microbiome analysis, and transcriptomic differential expression.
- Key packages:
mgcv(GAMs with thin-plate regression splines, REML optimization, AICc selection),survival(survfit for Kaplan-Meier),DESeq2(differential expression),GOATools,KOBAS(GO and KEGG enrichment, FDR-adjusted P ≤ 0.05).
- Key packages:
- IBM SPSS AMOS: Used for Structural Equation Modeling (SEM).
- SAS 9.4: Used for ANOVA, Kruskal-Wallis tests, and Generalized Estimating Equations (GEE).
- QIIME2: Used for 16S rRNA sequencing analysis (alpha/beta diversity, ANCOM, ANOVA, LEfSe).
- Excel 2019: Used for basic data organization and Shapiro-Wilk/F-tests for normality and homogeneity of variance.
Statistical Methods Summary:
- Normality & Variance: Shapiro-Wilk test and F-test.
- Nonlinear relationships: GAMs (thin-plate regression splines).
- Causal pathways: SEM.
- Repeated measures: GEE (with experimental group as independent variable and food intake as covariate).
- Group comparisons: One-way ANOVA (LSD for homogeneity, Dunnett T3 for heterogeneity), Kruskal-Wallis test, independent t-test.
- Microbiome: Alpha/beta diversity (Kruskal-Wallis, ANOSIM), Spearman correlation (immune factors vs. microbiota).
- Transcriptome: Z-score normalization for heatmaps (hierarchical clustering using Ward’s method and Euclidean distance).
Access information
Other publicly accessible locations of the data:
- Not applicable. All data generated during this study are included in this published Dryad dataset.
Data was derived from the following sources:
- Meteorological data were obtained from the China Meteorological Service Center (https://data.cma.cn/). Normalized Difference Vegetation Index (NDVI) data were obtained from the Moderate-resolution Imaging Spectroradiometer (MODIS) dataset provided by EarthData (https://search.earthdata.nasa.gov/search). All other data (rodent density, flea index, NDVI, experimental data) were generated by the authors.
