Mixed population trends inside a California protected area: Evidence from long-term community science monitoring
Data files
Dec 22, 2023 version files 91.54 KB
-
meta_analysis_results.csv
11.41 KB
-
README.md
15.26 KB
-
Species_Organization.csv
64.88 KB
Abstract
Protected areas are one of the most widespread and accepted conservation interventions, yet their population trends are rarely compared to regional trends to gain insight into their effectiveness. Here, we leverage two long-term community science datasets to demonstrate mixed effects of protected areas on long-term bird population trends. We analyzed 31 years of bird transect data recorded by community volunteers across all major habitats of Stanford University’s Jasper Ridge Biological Preserve to determine the population trends for a sample of 66 species. We found that nearly a third of species experienced long-term declines, and on average, all species declined by 12%. Further, we averaged species trends by conservation status and key life history attributes to identify correlates and possible drivers of these trends. Observed increases in some cavity-nesters and declines of scrub-associated species suggest that long-term fire suppression may be a key driver, reshaping bird communities through changes in forest and chaparral structure and composition. Additionally, we compared our results to those of the North American Breeding Bird Survey’s Central California Coast region (n = 55 species) to place Jasper Ridge in a broader context. Most species experienced similar directional population trends inside vs. outside of the preserve, and only eight species (14.5%) did better inside this small, protected area. Therefore, we must identify relevant management strategies for declining populations and explicitly consider how existing protected areas target and manage each species. Further, this analysis underscores the importance of local and national community science for revealing nuanced long-term bird population trends.
README: Data and model code from: Mixed population trends inside a California protected area: evidence from long-term community science monitoring
https://doi.org/10.5061/dryad.6t1g1jx2b
Here, we provide the R code used to model the abundance for each species in the Jasper Ridge Biological Preserve. We have also provided a spreadsheet with each species' life history traits, taxonomy, annual trends in the preserve, and annual trends in the surrounding region (BCR 32) from the North American Breeding Bird Survey. Finally, we have attached an R code that analyzes the trends for various life history traits and taxonomic families, compares trends within the protected area and in the surrounding region, and produces figures 2, 4, and 5 in the main manuscript and all supplementary material figures.
Description of the data and file structure
* *
The JRBP_Transect_Data_Species.R file provides the code required to create a generalized linear mixed model for each species in R-INLA and extract the percent change in abundance over the study period. Each model also includes the priors and distribution chosen for each species. Priors were selected individually for each species and adjusted so that the year and day-of-year random walks displayed an ecologically realistic amount of non-linearity. For example, in models for Neotropical migratory birds, the day-of-year random walks were expected to show a distinct peak in abundance during the April-July breeding season. The file also includes all code used to confirm or deny the validity of each model. While the raw Jasper Ridge bird transect dataset required to run this code is not publicly available, access can be granted to interested parties at https://jrbp.stanford.edu/research/data.
The Species_Organization.csv document contains a list of all species observed (without incidental observations), associated life history and taxonomic traits, conservation priority, level of endemism, Jasper Ridge trends in abundance with associated model details, and surrounding regional trends from the North American Breeding Bird Survey's BCR 32 California Central Coast region. Each row is a different species, while the explanation for each column is outlined below.
X: row names, from the species that was documented in the most surveys to the least.
Species_Code: four letter species codes prepared by the Institute for Bird Populations https://www.birdpop.org/docs/misc/Alpha_codes_eng.pdf.
Species_Common_Name: standardized by the American Birding Association; Check-List of North American Birds (Online). J. Ornithol.
Scientific_Name: standardized by the American Birding Association
Transects_Sighted: number of surveys the species was detected in, aka how common the species was in Jasper Ridge.
Individuals_Sighted: number of individuals of the species detected across all surveys in Jasper Ridge.
Taxonomic_Order: from most basal to least, retrieved from Wilman, Hamish, Jonathan Belmaker, Jennifer Simpson, Carolina de la Rosa, Marcelo M. Rivadeneira, and Walter Jetz. 2014. EltonTraits 1.0: Species-Level Foraging Attributes of the World’s Birds and Mammals. Ecology 95: 2027–2027.
Species_Level_Identification: 1 if identified to species level, 0 if not identified to species level.
PassNonPass: member of Passeriformes or not (Nonpasseriformes), NA if only identified as a bird.
FamilyLatin: Family the species belongs to, NA if only identified as a bird.
California.Near.Endemic.c: binary, over 80% of the species' range is in California.
West.Coast.Endemic.c: binary, over 80% of the species' range is along the West Coast.
No.Eastern.North.America.c: binary, this species is not a common resident or migrant east of the Mississippi River.
Diet.Inv: percent of the species diet that is invertebrate prey (Wilman et al. 2014).
ForStrat.ground: percent of time this species spends foraging on the ground (Wilman et al. 2014).
ForStrat.understory: percent of time this species spends foraging in the understory (Wilman et al. 2014).
ForStrat.midhigh: percent of time this species spends foraging in the mid/high canopy (Wilman et al. 2014).
ForStrat.canopy: percent of time this species spends foraging in the canopy (Wilman et al. 2014).
ForStrat.ground.weighted: equals ForStrat.ground*4 + ForStrat.understory*3 + ForStrat.midhigh*2 + ForStrat.canopy*1.
BodyMass.Value: body mass from (Wilman et al. 2014).
Body_mass_log: logarithmic body mass (Wilman et al. 2014).
Habitat.c: categorical preferred habitat from Powell, Hugh, Victoria Campbell, Gustave Axelson, Pat Leonard, and Miyoko Chu, eds. 2022. All About Birds. Ithaca, NY, USA: Cornell Laboratory of Ornithology; we only searched for values for species where we attempted models, NA if we did not search for a value.
Food.c: categorical diet from (Powell et al. 2022); we only searched for values for species where we attempted models, NA if we did not search for a value.
Nesting.c: categorical nesting location from (Powell et al. 2022); we only searched for values for species where we attempted models, NA if we did not search for a value.
Behavior.c: categorical foraging behavior from (Powell et al. 2022); we only searched for values for species where we attempted models, NA if we did not search for a value.
SF_Migratory.c: categorical migratory status in the San Francisco Bay Area around the study site from (Powell et al. 2022); we only searched for values for species where we attempted models, NA if we did not search for a value.
HWI: hand wing index, 100*Dk/Lw, where Dk is Kipp's distance (the distance between the tip of the first secondary feather and the tip of the longest primary feather) and Lw is wing length; the hand wing index (an estimate of wingtip pointedness) is widely adopted as a proxy for flight efficiency and dispersal in birds.
CA_Special_Concern: binary from Shuford, W.D. & Gardali, T. 2008. California Bird Species of Special Concern: A Ranked Assessment of Species, Subspecies, and Distinct Populations of Birds of Immediate Conservation Concern in California. Studies of Western Birds, Western Field Ornithologists, Camarillo, CA, USA and California Department of Fish and Game, Sacramento, CA, USA, 1.
CA_PIF_Focal_Species: binary, 1 if noted in CalPIF (California Partners in Flight) (2002). Version 2.0. The Oak Woodland Bird Conservation Plan: A Strategy for Protecting and Managing Oak Woodland Habitats and Associated Birds in California (S. Zack, lead author). Point Reyes Bird Observatory, Stinson Beach, CA, USA. http://www.prbo.org/calpif/plans.html; RHJV (Riparian Habitat Joint Venture) (2004). The riparian bird conservation plan: A strategy for reversing the decline of riparian associated birds in California, version 2.0. California Partners in Flight. http://www.prbo.org/calpif/pdfs/riparian_v-2.pdf; or CalPIF (California Partners in Flight) (2004). Version 2.0. The Coastal Scrub and Chaparral Bird Conservation Plan: A Strategy for Protecting and Managing Coastal Scrub and Chaparral Habitats and Associated Birds in California (J. Lovio, lead author). PRBO Conservation Science, Stinson Beach, CA, USA. http://www.prbo.org/calpif/plans.html. "
PIF: binary, 1 if noted in Will, T., Stanton, J.C., Rosenberg, K.V., Panjabi, A.O., Camfield, A., Shaw, A., Thogmartin, W.E., Blancher, P.J., Will, T. & Stanton, J.C. 2019. Handbook to the Partners in Flight Population Estimates Database Version 3.0. Partners in Flight Technical Series, no. 7.
U_value: specified for each species' R-INLA model, determines prior, NA indicates that the model failed validation or was not attempted if X > 100.
Distribution: Selected distribution for each species' R-INLA model, chosen to limit over- or under-dispersion, NA indicates that the model failed validation or was not attempted if X > 100.
Dispersion: a measure of the amount of variability in the data set based on the model; under dispersed models have a value less than 1 while over dispersed models are greater than 1, NA indicates that the model failed validation or was not attempted if X > 100.
WAIC: a generalized version of the Akaike Information Criterion (AIC), NA indicates that the model failed validation or was not attempted if X > 100.
DIC: Deviance Information Criterion is a hierarchical modelling generalization of the AIC, NA indicates that the model failed validation or was not attempted if X > 100.
Credible_Trend: binary, 1 if credible, 0 if not, NA indicates that the model failed validation or was not attempted if X > 100.
Population_Trend_Magnitude: percent population change over the 31 year study, NA indicates that the model failed validation or was not attempted if X > 100.
PTM_LowerCI: lower credible interval of population trend magnitude, NA indicates that the model failed validation or was not attempted if X > 100.
PTM_UpperCI: upper credible interval of population trend magnitude, NA indicates that the model failed validation or was not attempted if X > 100.
PTM_Credible_Interval: credible interval of population trend magnitude, NA indicates that the model failed validation or was not attempted if X > 100.
PTM_Standard_Error: PTM_Credible_Interval divided by 1.96, NA indicates that the model failed validation or was not attempted if X > 100.
PTM_Variance: square of PTM_Standard_Error, NA indicates that the model failed validation or was not attempted if X > 100.
PTM_Average_Annual_Trend: equals (1 + Population_Trend_Magnitude)^(1/31 years) - 1, NA indicates that the model failed validation or was not attempted if X > 100.
AAT_LowerCI: equals (1 + PTM_LowerCI)^(1/31 years) - 1, NA indicates that the model failed validation or was not attempted if X > 100.
AAT_UpperCI: equals (1 + PTM_UpperCI)^(1/31 years) - 1, NA indicates that the model failed validation or was not attempted if X > 100.
AAT_Credible_Interval: the average distance from the average annual trend to AAT_LowerCI and AAT_UpperCI, NA indicates that the model failed validation or was not attempted if X > 100.
AAT_Standard_Error: equals AAT_Credible_Interval/1.96, NA indicates that the model failed validation or was not attempted if X > 100.
AAT_Variance: square of AAT_Standard_Error, NA indicates that the model failed validation or was not attempted if X > 100.
S32_C: credibility of North American Breeding Bird Survey (BBS) trend from the BCR_32 California Coast region, 1 blue on webpage, 2 indicates yellow, and 3 indicates red, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
BBS_ERROR: 1 if there was an error in loading, 0 if none, excluded from analysis, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
BBS_Notes: notes on trend, often errors, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_CredTrend: binary, 1 if credible, 0 if not, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_Trend: BBS average annual trend in the BCR_32 region between 1989 and 2019, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_LowerCI: lower credible interval of the BBS average annual trend in the BCR_32 region between 1989 and 2019, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_UpperCI: upper credible interval of the BBS average annual trend in the BCR_32 region between 1989 and 2019, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_Credible_Interval: average distance from the S32_Trend to S32_LowerCI or S32_UpperCI, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_Standard_Error: S32_Credible_Interval/1.96, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
S32_Variance: square of S32_Credible_Interval, NA indicates that the Jasper Ridge model failed validation, was not attempted if X > 100, or was not available.
The BBS_and_JRBP_Comparison.R file contains all code used to analyze the trends for various life history traits and taxonomic families, compare trends within the protected area and in the surrounding region, and produce figures 2, 4, and 5 in the main manuscript and all supplementary material figures. Please put all files into a folder named "Data_JRBP_Bird_Transect" on your desktop for the file paths in the R code to work correctly.
Finally, the meta_analysis_results.csv document contains the results of many random effects models (REM) conducted in BBS_and_JRBP_Comparison.R. Each row is a different grouping of species, while the explanation for each column is outlined below.
Group: For each REM, we restricted the species to those that shared this attribute (life history trait, endemism, taxonomy, conservation status, etc.).
YesFigure: binary, notates if the group's averaged trend was included in Figure 4.
FigureOrder: indicates the order of many REM results in Figure 4; NA indicates that the REM’s results were not included in the figure.
JRBPonly: binary, TRUE if the species’ population trends fed into the REM are solely from Jasper Ridge; FALSE if the trends are the difference between the Jasper Ridge and BBS BCR_32 region.
n: how many species trends were fed into REM; also the number of species that share that attribute and have valid models for Jasper Ridge or both Jasper Ridge and the BBS BCR_32 region.
beta: provides the group's average annual trend from the REM.
se: standard error from the REM.
zval: z-value from the REM.
pval: p-value from the REM
ci.lb: lower bound of the 95% confidence interval from the REM.
ci.ub: upper bound of the 95% confidence interval from the REM.
tau2: the τ2 from the REM, the extent of variation among the effects observed in different population trends (between-trend variance), estimated using the restricted maximum-likelihood estimator.
se.tau2: standard error of τ2 from the REM.
I2: I2 statistic from the REM (total heterogeneity / total variability); see Higgins, J. P. T., & Thompson, S. G. (2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21(11), 1539–1558. https://doi.org/10.1002/sim.1186 for more detail.
H2: H2 statistic from the REM (total variability / sampling variability).
QE: results of the Q-test for heterogeneity, a test to quantify the level of heterogeneity among the trends fed into the REM.
QEp: p-value from the Q-test for heterogeneity; significance notes that the trends fed into the REM are heterogeneous.
mag: the group's average percent increase or decrease between 1989 and 2020.
Sharing/Access information
* *
Access to the raw Jasper Ridge bird transect dataset is not publicly available, but it can be granted to interested parties: https://jrbp.stanford.edu/research/data
Analyzed results from the North American Breeding Bird Survey Dataset can be found here: https://www.mbr-pwrc.usgs.gov
Methods
From 1989 to 2020, volunteer observers conducted monthly surveys of six sectors within Stanford University's Jasper Ridge Biological Preserve (JRBP). Each survey consisted of a trail-based transect in which a group of observers walked the trail in the morning and counted all birds detected over roughly 3 hours. Observers recorded the number of each species seen or heard along the route, regardless of the distance to the bird. Over 31 years of surveys, 192 observers conducted 2,055 transects and recorded a total of 473,401 observations of 184 species (91% of JRBP’s documented avian richness). We used these data to estimate long-term avian population trends at JRBP. Prior to analy- sis, we performed extensive data cleaning, including the standardization of species names and observer identity. Unlikely species without notes or supporting information were removed from the analysis. All transects with fewer than seven species (n = 30) were considered incidental and removed. These transects were often performed during suboptimal conditions (e.g. wind or rain) and/or were of abnormally short duration. Finally, we limited our analysis to the 100 most consistently detected species (those detected in the greatest number of transects).