Ecology of community reassembly: Movements and diets of megafauna during a decade of restoration in Mozambique

Hutchinson, Matthew 1 ; Walker, Hallie2; Abraham, Joel2; Becker, Justine3; Gonçalves, Dominique4; Nutter, Ciara2; Pansu, Johan5; Phillips, Erin2; Potter, Arjun6; Tenente, Beto7; Stalmans, Marc7; Long, Ryan8; Pringle, Robert2

Research facility: Gorongosa National Park

Published Feb 11, 2026 on Dryad. https://doi.org/10.5061/dryad.m905qfv85

Abstract

This dataset documents large-herbivore movement, morphology, condition, fate, and diet during community reassembly in Gorongosa National Park, Mozambique (2013–2023), following megafaunal declines during civil war and subsequent restoration, including apex predator reintroductions beginning in 2018. The study period spans substantial interannual climatic variability, including extreme wet and dry years.

The movement dataset comprises GPS telemetry records from 277 individuals across seven herbivore species (Cape bushbuck, nyala, greater kudu, common eland, waterbuck, plains zebra, and African elephant), with median monitoring durations ranging from 13 to 707 days and median location counts per individual ranging from 1,171 to 33,122 fixes. For 302 immobilized individuals, we provide morphological measurements (chest girth, body length, hind-foot length, body mass), reproductive and nutritional condition metrics (ultrasound measurements, palpation scores), and fate data (mortality date and cause, when known).

The diet dataset includes DNA metabarcoding results from 3,785 fecal samples collected from 27 mammal species (11 families, 7 orders), identifying 516 food-plant taxa representing at least 87 families and 39 orders. Sample sizes per species range from 1 to 499 (median = 92), with larger sample sizes for dominant large herbivores (median = 216).

All datasets include associated field metadata (date, time, location, sex, age), laboratory notes, and plant taxonomic information. In addition, filtering scripts and raw data are provided, facilitating alternative filtering approaches. These data support analyses of movement ecology, trophic interactions, nutritional ecology, and large-mammal community reassembly under restoration and climatic variability.

Access this dataset on Dryad: DOI: 10.5061/dryad.m905qfv85

Data are associated with: Hutchinson et al. 2026. Ecology. Please see publication for full details of methodlogy, study design, and data use.

This repository contains long-term ecological data collected in Gorongosa National Park, Mozambique. The dataset includes historical and contemporary rainfall data, GPS telemetry data from seven herbivore species, capture and morphological measurements, dietary DNA metabarcoding data (processed and raw), plant barcode taxonomy assignments, primer diagnostics, and complete bioinformatic filtering scripts.

These files document environmental variation, animal movement, demography, nutritional condition, and trophic interactions during megafaunal recovery.

Description of the data and file structure

Unless otherwise noted:

Missing values are coded as NA.
Spatial coordinates are WGS1984 (decimal degrees).
GPS timestamps are local time (UTC+02:00).
Dates follow ISO format (YYYY-MM-DD).
All data (points 1-9) described below are present in the .ZIP file "R1_DataS1_Hutchinson_etal_Ecology.zip" except points 7 and 8, which are provided as individual files here.

1. Climate Data

annual_rainfall_1957-2024.csv

Annual rainfall totals (mm) for Chitengo Camp, 1957–2024 (39 rows).

Columns:

Year
Rainfall_mm

Anomalies:

Missing years: 1970–1998, 2010–2011.
Incomplete coverage: 2005 (missing 1 month), 2009 (missing 4 months).
2008 total (204 mm) deemed unreliable.
Data from 2013–2023 are reliable; from 2015 onward measured via both automatic and manual gauges.

monthly_rainfall_2011-2024.csv

Monthly rainfall totals (mm), October 2011–December 2024 (159 rows).

Columns:

Year
Month
Rainfall_mm

No known anomalies.

2. Capture, Collar, and Fate Data

COLLAR-CAPTURE-AND-FATE-RECORDS_2014-2023.csv

302 rows; each row = one capture event.

Data columns are:

AnimalID Unique individual identifier encoding species code, year of collar deployment, sex, collar serial number, and deployment number. Deployment number indicates sequence of collar use within a year.
Species Scientific name of the species.
Sex Sex of the individual. Possible values: f (female), m (male)
CollarSerial Manufacturer serial number of the GPS collar.
Latitude Latitude of capture location (decimal degrees, WGS1984).
Longitude Longitude of capture location (decimal degrees, WGS1984).
DeployDate Date of capture and collar deployment (YYYY-MM-DD).
DeployYear Year of capture (YYYY).
DeployMonth Month of capture (written in long form, e.g., “July”).
DeployDayOfYear Day of year of capture (1–366).
FirstFixDate Date of first GPS location fix recorded by the collar (YYYY-MM-DD).
LastFixDate Date of last GPS location fix recorded by the collar (YYYY-MM-DD).
Fetus Presence of fetus determined by ultrasonography. Possible values: 1 (present), 0 (absent)
Lactate Lactation status at capture. Possible values: 1 (lactating), 0 (not lactating)
MaxFat_mm Maximum rump fat depth (mm), measured via ultrasonography.
Bfemoris_mm Thickness of biceps femoris muscle (mm), measured via ultrasonography.
Ldorsi_mm Thickness of longissimus dorsi muscle (mm), measured via ultrasonography.
SSligament_score Sacrosciatic ligament palpation score (0–6; ordinal).
Lumbar_score Lumbar vertebrae palpation score (0–6; ordinal).
Sacrum_score Sacrum palpation score (0–6; ordinal).
BaseTail_score Base-of-tail palpation score (0–6; ordinal).
Caudal_score Caudal vertebrae palpation score (0–6; ordinal).
ChestGirth_cm Chest circumference at the withers (cm).
BodyLength_cm Dorsal length from nose to base of tail (cm).
HindFoot_cm Hind foot length from hoof to hock (cm).
MeasuredWeight_kg Body mass measured directly in the field using a portable scale (kg).
EstimatedWeight_kg Body mass estimated (kg) using the regression: Mass (kg) = (1.6036 × ChestGirth – 78.681) – 5.35 Estimated values are also provided for individuals that were directly weighed.
FecalSample Indicator that a fecal sample was collected and produced usable dietary data. Possible values: 1 (yes), 0 (no)
FecalDietID Identifier linking the individual to fecal samples in diet datasets (step8_final_dataset_rarefied_RRA.csv and step8_final_dataset_raw_reads.csv). NA if none.
HairSample Indicator that a hair sample was collected. Possible values: 1 (yes), 0 (no)
BloodSample Indicator that a blood sample was collected. Possible values: 1 (yes), 0 (no)
HornSample Indicator that a horn sample was collected. Possible values: 1 (yes), 0 (no)
Mortality Indicator that the individual died while the GPS collar was active. Possible values: 1 (died), 0 (survived or collar ceased transmitting prior to death)
MortDate Estimated date of mortality (YYYY-MM-DD). NA if not applicable.
MortCause Confirmed cause of death (text field). NA if unknown or not applicable.
OtherNotes Free-text notes documenting measurement uncertainty, qualitative condition or reproductive observations, capture complications, or interpretive cautions.

3. GPS Collar Location Data

Each file contains time-stamped GPS fixes with DOP accuracy metrics.

Common columns:

AnimalID
Species
Timestamp
Year
DayOfYear
Latitude
Longitude
HDOP Horizontal dilution of precision: uncertainty in latitude and longitude driven by satellite position at time of fix
DOP Dilution of precision: uncertainily of 3D location (lat, long, elevation) driven by satellite position at time of fix

Files:

GPS-COLLAR-LOCATION-RECORDS_BUSHBUCK.csv (656,828 records; 103 individuals)
GPS-COLLAR-LOCATION-RECORDS_ELAND.csv (131,173 records; 10 individuals)
GPS-COLLAR-LOCATION-RECORDS_ELEPHANT.csv (839,947 records; 18 individuals)
GPS-COLLAR-LOCATION-RECORDS_KUDU.csv (1,299,250 records; 80 individuals)
GPS-COLLAR-LOCATION-RECORDS_NYALA.csv (444,809 records; 37 individuals)
GPS-COLLAR-LOCATION-RECORDS_WATERBUCK.csv (61,744 records; 22 individuals)
GPS-COLLAR-LOCATION-RECORDS_ZEBRA.csv (39,118 records; 7 individuals)

4. Processed Diet Metabarcoding Data

step8_final_dataset_rarefied_RRA.csv

3,785 samples × 532 columns.

Columns 1–16: metadata
Columns 17–532: plant mOTUs (Relative Read Abundance; rarefied to 979 reads per sample)

step8_final_dataset_raw_reads.csv

Same structure as above.

Plant mOTUs reported as mean read counts across PCR replicates (unrarefied).

step8_final_metadata.csv

Metadata-only file (3,785 rows; 17 columns).

Includes:

Sample identifiers
Species (field and genetically confirmed)
Sex
Collection date/time
Latitude/Longitude
Collection confidence score
AnimalID
Mean read depth

5. Plant Barcode Taxonomy

step8_final_mOTU_taxonomy.csv

516 mOTUs × 17 columns.

Includes:

DNA sequence
Best database match
NCBI taxid
Taxonomic rank
Post-hoc resolution fields
CombinedID (matches plant columns in diet datasets)

step7_resolved_mOTUs.csv

145 most common mOTUs (>90% of reads).

Includes:

BLAST matches
Mozambique and Gorongosa filtering
FinalID and TidyFinalID

mOTU_matching_Potter2022.csv

Links current mOTU taxonomy to Potter et al. (2022) plant trait assignments.

Includes:

Sequence
CombinedID
Resolved_SpeciesList
PotterSpeciesID fields

6. Primer Diagnostics

primer_template_alignments.csv

516 mOTUs × 44 columns.

Reports:

Primer matches
Total mismatches per primer
Position-specific mismatch indicators (F1–F17; R1–R22; binary 0/1)

7. Raw DNA Metabarcoding Data – Diet (trnL)

Gzipped FASTQ files (paired-end aligned and sample-assigned via OBITools v1.2.11):

png13_jul_lib1_assigned.fastq.gz
png13_jul_lib2_assigned.fastq.gz
png14_jul_assigned.fastq.gz
png15_jul_lib1_assembled_assigned.fastq.gz
png15_jul_lib2_assembled_assigned.fastq.gz
png15_jul_lib3_assembled_assigned.fastq.gz
png15_jul_lib4_assembled_assigned.fastq.gz
png15_jul_lib5_assembled_assigned.fastq.gz
png15_jul_part2_assigned.fastq.gz
png16_jul_lib1_assembled_assigned.fastq.gz
png16_jul_lib2_assembled_assigned.fastq.gz
png16_jul_lib3_assembled_assigned.fastq.gz
png17_jul_assembled_assigned.fastq.gz
png17_julRedo_nov_assembled_assigned.fastq.gz
png18_apr_assembled_assigned.fastq.gz
png18_jul_lib1_assembled_assigned.fastq.gz
png18_jul_lib2_assembled_assigned.fastq.gz
png18_jul_lib3_assembled_assigned.fastq.gz
png18_nov_assembled_assigned.fastq.gz
png19_apr_assembled_assigned.fastq.gz
png19_jul_lib1_assembled_assigned.fastq.gz
png19_jul_lib2_assembled_assigned.fastq.gz
png19_jul_lib3_assembled_assigned.fastq.gz
png19_nov_assembled_assigned.fastq.gz
png20_apr_assembled_assigned.fastq.gz
png21_jul_lib1_assembled_assigned.fastq.gz
png21_jul_lib2_assembled_assigned.fastq.gz
png21_jul_lib3_assembled_assigned.fastq.gz
png21_22_lib1_assembled_assigned.fastq.gz
png21_22_lib2_assembled_assigned.fastq.gz
png21_22_lib3_assembled_assigned.fastq.gz

Each FASTQ file includes:

Sample identifier
Sequencing library
Alignment metrics
DNA sequence
Phred +33 quality scores

8. Raw DNA Metabarcoding Data – Host Identification (16S)

FASTQ files:

sample_id_test_lib1_assembled_assigned.fastq.gz
sample_id_test_lib2_assembled_assigned.fastq.gz
sample_id_test_2022_lib1_assembled_assigned.fastq.gz
sample_id_test_2023_lib1_assembled_assigned.fastq.gz
sample_id_test_2023_lib2_assembled_assigned.fastq.gz

Same structure as trnL FASTQ files; used for host species confirmation.

9. Filtering Scripts and Supporting Files

Directory structure:

GPS-collar-location-data-processing.R
filtering_prerequisites.txt
bioinformatics_step1/
bioinformatics_step2/
ngsfilter_files/
metadata_files/
additional_files_scripts/

These files reproduce the full filtering workflow from raw FASTQ data (Sections 7 & 8) to final processed datasets (Sections 4 & 5). Includes R scripts, BASH scripts, OBITools configuration files, FASTA references, and intermediate objects.

Scripts are fully commented and intended for users wishing to refilter raw data.

Sharing/Access Information

All data are original to long-term monitoring in Gorongosa National Park.

Users should cite the associated publication when using these data.

Code/Software

OBITools v1.2.11 used for paired-end alignment and sequence assignment.
R Statistical Software used for downstream filtering and final dataset construction.
FASTQ files are gzip-compressed.
CSV files are compatible with R, Python, MATLAB, Julia, and Excel.

Filtering scripts are provided to enable complete reproducibility of the diet metabarcoding workflow.

Ecology of community reassembly: Movements and diets of megafauna during a decade of restoration in Mozambique

Data files

Abstract

README: Long-term movement, climate, morphology, and diet metabarcoding data from Gorongosa National Park (2013–2023)

Description of the data and file structure

1. Climate Data

annual_rainfall_1957-2024.csv

monthly_rainfall_2011-2024.csv

2. Capture, Collar, and Fate Data

COLLAR-CAPTURE-AND-FATE-RECORDS_2014-2023.csv

3. GPS Collar Location Data

4. Processed Diet Metabarcoding Data

step8_final_dataset_rarefied_RRA.csv

step8_final_dataset_raw_reads.csv

step8_final_metadata.csv

5. Plant Barcode Taxonomy

step8_final_mOTU_taxonomy.csv

step7_resolved_mOTUs.csv

mOTU_matching_Potter2022.csv

6. Primer Diagnostics

primer_template_alignments.csv

7. Raw DNA Metabarcoding Data – Diet (trnL)

8. Raw DNA Metabarcoding Data – Host Identification (16S)

9. Filtering Scripts and Supporting Files

Sharing/Access Information

Code/Software