Arthropod density observations and measures of crop production in California citrus
Data files
Apr 28, 2025 version files 1.41 MB
-
Citrusformatics_dataset_anonymized.csv
1.41 MB
-
README.md
6.58 KB
Abstract
Early studies and theory suggested that complex landscapes harboring remnants of natural land should support natural enemy populations and reduce pest buildup in adjacent crops, whereas landscapes interspersed with urban land often provide alternate host plants of crop pests, facilitating pest spillover and amplifying pest pressure. However, recent meta-analyses have demonstrated that both pest and beneficial agricultural arthropods respond inconsistently to surrounding landscapes. These meta-analyses relied on studies of one to two pests per crop across many different crop and landscape contexts, which limits inferences about how growers might design landscapes for effective control of a full suite of pests attacking a given crop. Here, we harnessed an ecoinformatics dataset from California Citrus to examine the effects of surrounding natural and urban land on the densities of a complete suite of seven major pest species (7,625 observations) and one beneficial predator (335 observations). We also explored landscape effects on pesticide use and fruit production. Despite restricting this analysis to data collected in the same region and cropping system, we found that arthropods still exhibited mixed responses to surrounding landscapes. Among the eight Citrus-associated arthropods surveyed, greater amounts of nearby natural land resulted in two outcomes beneficial for farmers (lower pest densities or fewer pesticide applications targeting that pest), three adverse outcomes, and three neutral outcomes. Similarly, greater amounts of urban land resulted in one beneficial outcome, three adverse outcomes, and four neutral outcomes for growers. Our economic analysis demonstrated no significant effects of either natural or urban land on total pesticide use or fruit quality, but reduced total fruit yield in Citrus groves with more nearby natural land. Taken altogether, our results do not demonstrate clear support for the retention of natural habitat or minimization of urban land near cropland solely for the purpose of enhancing conservation biocontrol. Nonetheless, the value of natural land extends far beyond its utility for conservation biocontrol, and agricultural landscapes must still be managed to strike a balance between crop production and the preservation of biodiversity and ecosystem function.
GENERAL INFORMATION
Title of Dataset: Arthropod density observations and measures of crop production in California Citrus
Author/Principal Investigator Information
Name: Mia Lippey
ORCID:0000-0002-8292-3444
Institution: University of California, Davis
Address: 1 Shields Ave, Davis, CA 95616
Email: mklippey@ucdavis.edu
Date of data collection: 2006 - 2019
Geographic location of data collection: San Joaquin Valley, California
Information about funding sources that supported the collection of the data: USDA-FACT
SHARING/ACCESS INFORMATION
Licenses/restrictions placed on the data: none; data were previously restricted, but all sensitive information has been redacted
Was data derived from another source?
If yes, list source(s): Citrusformatics database (privately hosted)
Recommended citation for this dataset: pending
DATA & FILE OVERVIEW
Are there multiple versions of the dataset? Yes
If yes, name of file(s) that was updated: Data_Dump_9_20_23
Why was the file updated? It was raw data that needed to be wrangled for analysis
When was the file updated? 09-27-2023
METHODOLOGICAL INFORMATION
Description of methods used for collection/generation of data: This data is an ecoinformatics dataset, so all arthropod observations and measures of crop production were reported by independent Citrus growers and associated pest control advisors in the San Joaquin Valley. Arthropod density observations were recorded by field scouts employed by growers, and each species was sampled a different way. Methods include presence/absence sampling and count data (more details described in associated publication). Measures of crop production were recorded by packing houses employed by growers, and they include metrics such as crop yield and crop quality. Pesticide use data were reported by pest control advisors employed by growers.
Methods for processing the data: The submitted data were processed from the raw data for various purposes. The first process was cleaning typos in the dataset. The second process included data interpolation for missing time steps or sites. Both were conducted by Bodil Cass, and specifics of the data wrangling process have been published.
Instrument- or software-specific information needed to interpret the data: The version of R used for this analysis is R version 4.3.2 (2023-10-31 ucrt). The packages need to run the script include tidyverse, dplyr, car, ape, mgcv, ggplot2, mdthemes, glmmTMB, DHARMa, and ggeffects.
Standards and calibration information, if appropriate: N/A
Environmental/experimental conditions: California has a Mediterranean climate. Summers are hot and dry, winters are mild and wet. Annual precipitation falls almost entirely in the winter months.
Describe any quality-assurance procedures performed on the data: Each step of data cleaning was checked by hand to ensure that mutations and various calculations were correct.
People involved with sample collection, processing, analysis and/or submission: Bodil Cass, Jay Rosenheim, and Mia Lippey.
DATA-SPECIFIC INFORMATION FOR: Arthropod density observations and measures of crop production in California Citrus
Number of variables: 43
Number of cases/rows: 3998
Variable List:
-
field_code: a number for each unique field site
-
crop_year -the year that the data were collected
-
field_year is a variable that pastes field_code with crop_year to create a unique identifier for each field site for each year
-
field_size_ha -size of focal citrus field in ha
-
observer -the anonymized name of pest control agents who collected the data. We aggregated individual PCAs into “PCA_group_1” and “PCA_group_2”, putting together PCAs with comparable sampling methods.
-
region_16 is a variable used to control for spatial autocorrelation which divides the entire San Joaquin Valley site into 16 equally sized squares (a proxy for lat/long, which is sensitive information and has been redacted)
-
tree_age -age of each citrus focal field
-
urban_proportions -the absolute proportional area of urban land within each circular buffer around each focal field at radii scales of 500m, 1km, 2km, and 4km
-
natural_proportions -the absolute proportional area of natural land within each circular buffer around each focal field at radii scales of 500m, 1km, 2km, and 4km
-
citrus_proportions -the absolute proportional area of citrus crops within each circular buffer around each focal field at radii scales of 500m, 1km, 2km, and 4km
-
other_cropland_proportions -the absolute proportional area of non-citrus cropland within each circular buffer around each focal field at radii scales of 500m, 1km, 2km, and 4km
-
other_noncropland_proportions -the absolute proportional area of other land uses (water, barren land, etc.) within each circular buffer around each focal field at radii scales of 500m, 1km, 2km, and 4km
-
katydid_area_percent_infested -an average of density observations for katydids per field per year
-
katydid_sprays -number of pesticide sprays targeting katydids per field per year
-
citricola_area_percent_infested -an average of density observations for citricola per field per year
-
citricola_sprays -number of pesticide sprays targeting citricola per field per year
-
fruit_dam_per_bin_red_scale -an average of density observations for red scale per field per year
-
red_scale_sprays -number of pesticide sprays targeting red scale per field per year
-
thrips_area_percent_infested -an average of density observations for thrips per field per year
-
thrips_sprays -number of pesticide sprays targeting thrips per field per year
-
red_mite_area_percent_infested -an average of density observations for red mite per field per year
-
red_mite_sprays -number of pesticide sprays targeting red mite per field per year
-
cottony_area_percent_infested -an average of density observations for cottony cushion scale per field per year
-
fruit_dam_per_bin_peelminer -an average of density observations for peelminer per field per year
-
Euseius_area_insect_count -an average of density observations for Euseius per field per year
-
total_num_sprays -total number of pesticide sprays across all pests per field per year
-
fruit_quality_metric -fruit quality of citrus blocks per year, on a continuous scale from 0-2
-
harvest_total_kg_per_ha -total fruit harvested per block per year in kg/ha
Missing data codes: N/A
Specialized formats or other abbreviations used: N/A
These data were aggregated from several independent sources. These include Citrus growers, pest control advisors, packing houses, the USGS National Land Cover Database. and the USDA CropScape data layer.
Natural, urban, and crop land use data were extracted from the USGS National Land Cover Database by creating circular buffers around focal Citrus fields at four spatial scales (500m, 1km, 2km, and 4km). These data were grouped into three main categories including urban land, natural land, cropland, and other noncropland. The "other" category comprised mostly water and barren land.
For citrus cover, we used the USDA CropScape Data Layer (CDL). We subtracted citrus cover from the NLCD cropland area to create the two land use categories of citrus and other cropland.
Arthropod observations and pesticide applications were reported by independent pest control advisors hired by Citrus growers. Each species was sampled differently, and the methods are described in detail in the associated manuscript.
Fruit quality and yield were reported by packing houses hired by Citrus growers.
Abundance observations of arthropods were averaged across samples taken within focal Citrus fields because fields were sampled more than once per year.
The unique identifier for each field is "field_code", and the unique identifier of the field per year is "field_year".