High resolution gridded climate data from TerraClimate (https://www.climatologylab.org/terraclimate.html) was downloaded from 1958-2022 for use in the Pacific Southwest Region Broader scale Monitoring Strategy. Total yearly Snow Water Equivalent data was summarized at various spatial scales (Forest, Ecological Province, and Regional level). An additional "Trend Data" column was calculated to provide a linear regression line within the Broader Scale Monitoring Strategy Climate Change Dashboard.

High-resolution reanalysis products could improve the representativeness of rainfall on high Andean basins, but their performance must be locally validated. We addressed the performance, accuracy, and ability of TerraClimate and CHIRPSv.2 to represent 36 years of rainfall (1985–2020) from 23 stations in the Upper Chicamocha River, a Colombian basin of complex terrain and tropical hydrometeorology. Using several statistical metrics at monthly, seasonal and annual scales, we found how both datasets overestimate rainfall as a function of elevation, with better performance and accuracy from CHIRPS (r ∼0.76, R2 ∼0.58, NSE ∼0.56, and low RMSE ∼33.7 mm/month, MAE ∼25.2 mm/month, ME ∼6.4 mm/month, and PBIAS ∼9.3), while TerraClimate overestimates inter-annual variability, especially between June and August. Seasonally, the datasets exhibit different spatial patterns and magnitudes, even after bias correction. The findings highlight the potential use and challenges of high-resolution datasets in basins with similar topography and hydrometeorology in the Andean region.

TerraClimate, küresel kara yüzeyleri için aylık iklim ve iklimsel su dengesi verilerinden oluşan bir veri kümesidir. Bu model, WorldClim veri setinden alınan yüksek mekansal çözünürlüklü klimatolojik normaller ile CRU Ts4.0 ve Japonya'nın 55 yıllık yeniden analizi (JRA55) verilerinden alınan daha düşük mekansal çözünürlüklü ancak zamana göre değişen verileri birleştiren, iklimsel olarak desteklenen bir enterpolasyon kullanır. Kavramsal olarak, prosedür, enterpolasyonlu …

IDN_CLI_SPEI_blend_0p042_1961_2021 is currently the only comprehensive high resolution Indonesia gridded historical dataset of SPEI blend and available for public.

The SPEI - https://spei.csic.es/ is an extension of the widely used SPI. The SPEI is designed to take into account both precipitation and potential evapotranspiration (PET) in determining drought. Thus, unlike the SPI, the SPEI captures the main impact of increased temperatures on water demand.

The IDN_CLI_SPEI_blend_0p042_1960_2021 is derived using precipitation and potential evapotranspiration from TerraClimate data - https://www.climatologylab.org/terraclimate.html, it has 0.042 degree gridded resolution, a monthly and available from 1958 to 2021. The calibration period is January 1961 to December 2020. The starting date of the dataset is 1960 in order to provide common information across the different SPEI time-scales.

The SPEI blend integrate several SPEI scales into a single product, combine 3-, 6-, 9-, 12- and 24-month SPEI to estimate the overall dry/wet condition.

The SPEI processed using climate_indices, an open source Python library providing reference implementations of commonly used climate indices. https://pypi.org/project/climate-indices/

Data was obtained from TerraClimate climate principal components analysis, which are the product used to assess climate deviation under +2C and +4C warming scenarios. We elected to use the principal component analysis with the top 5, unique components, with reclassified values. Original values were on a 0- 1 scale, but were rescaled to 0- 100 to conform to other model variables.

TerraClimate es un conjunto de datos del clima mensual y el balance hídrico climático para las superficies terrestres globales. Utiliza la interpolación asistida por el clima, que combina las normales climatológicas de alta resolución espacial del conjunto de datos de WorldClim con datos de menor resolución espacial, pero que varían con el tiempo, del CRU Ts4.0 y el Reanálisis japonés de 55 años (JRA55). Conceptualmente, el procedimiento aplica anomalías interpoladas que varían con el tiempo de CRU Ts4.0/JRA55 a la climatología de alta resolución espacial de WorldClim para crear un conjunto de datos de alta resolución espacial que abarca un registro temporal más amplio. La información temporal se hereda de CRU Ts4.0 para la mayoría de las superficies terrestres globales en cuanto a temperatura, precipitación y presión de vapor. Sin embargo, los datos de JRA55 se utilizan para las regiones en las que los datos de CRU no tenían estaciones climáticas que contribuyeran (incluida toda la Antártida y partes de África, América del Sur e islas dispersas). Para las variables climáticas principales de temperatura, presión de vapor y precipitación, la Universidad de Idaho proporciona datos adicionales sobre la cantidad de estaciones (entre 0 y 8) que contribuyeron a los datos de CRU Ts4.0 que utiliza TerraClimate. JRA55 se usó exclusivamente para la radiación solar y las velocidades del viento. Además, TerraClimate produce conjuntos de datos mensuales del balance hídrico superficial con un modelo de balance hídrico que incorpora la evapotranspiración de referencia, la precipitación, la temperatura y la capacidad de agua disponible en el suelo interpolada. Se utilizó un modelo climático modificado de balance hídrico de Thornthwaite-Mather y datos de capacidad de almacenamiento de agua del suelo extraíble en una cuadrícula de 0.5° de Wang-Erlandsson et al. (2016). Limitaciones de los datos: Las tendencias a largo plazo en los datos se heredan de los conjuntos de datos principales. TerraClimate no se debe usar directamente para evaluaciones independientes de las tendencias. TerraClimate no capturará la variabilidad temporal en escalas más finas que los conjuntos de datos principales y, por lo tanto, no puede capturar la variabilidad en las inversiones y las proporciones de precipitación orográfica. El modelo de balance hídrico es muy simple y no tiene en cuenta la heterogeneidad en los tipos de vegetación ni su respuesta fisiológica a las condiciones ambientales cambiantes. Validación limitada en regiones con pocos datos (p.ej., Antártida).

This dataset contains 14 parquet-format files with monthly data.

    File
    Indicator
    Unit


    aet.parquet
    Actual Evapotranspiration
    mm


    def.parquet
    Climate Water Deficit
    mm


    pdsi.parquet
    Palmer Drought Severity Index (PDSI)
    unitless


    pet.parquet
    Precipitation
    mm


    ppt.parquet
    Potential evapotranspiration
    mm


    q.parquet
    Runoff
    mm


    soil.parquet
    Soil Moisture
    mm


    srad.parquet
    Downward surface shortwave radiation
    W/m2


    swe.parquet
    Snow water equivalent
    mm


    tmax.parquet
    Maximun Temperature
    °C


    tmin.parquet
    Minimum Temperature
    °C


    vap.parquet
    Vapor pressure
    kPa


    vpd.parquet
    Vapor Pressure Deficit
    kpq


    ws.parquet
    Wind speed
    m/s

TerraClimate est un ensemble de données mensuelles sur le climat et l'équilibre hydrique climatique pour les surfaces terrestres mondiales. Elle utilise l'interpolation assistée par le climat, en combinant les normales climatologiques à haute résolution spatiale de l'ensemble de données WorldClim avec des données à résolution spatiale plus grossière, mais variant dans le temps, provenant de CRU Ts4.0 et de la réanalyse japonaise sur 55 ans (JRA55). Conceptuellement, la procédure applique une interpolation …

We update the total runoff and runoff partitioned from snowmelt from TerraClimate using a modified water balance model that better accounts for baseflow lags in hydrology for use in basin level water supply availability at monthly time scales. These data complement existing TerraClimate 1-d water balance similations, but show much has improved montly correlative skill (median r=0.88) to observed streamflow across watersheds globally watersheds compared with the original method from TerraClimate (median r=0.75). These data can be used to quantify gaps in monthly water demand for downstream water uses including agriculture., The original TerraClimate (Abatzoglou et al., 2018) runoff data (q) assumed that all water (rain plus snowmelt, P) in excess of the additional soil water holding capacity (maximum soil water storage minus soil moisture or SR) and monthly evapotranspiration (ET) – or water surplus – immediately runs off that month. S = max(0, 0.95*P - ET-SR) q = 0.05*P + S While this water balance approach may be valid at local scales, the approach does not account for a variety of factors including baseflow contributions or stream routing and transit time. Some studies show that the transit time for runoff to reach a basin terminus is up to a month (Allen et al., 2018). We updated the approach for simulating runoff to account for the contributions of baseflow and transit time following Wolock and McCabe (1999). Specifically, half of the surplus generated each month becomes runoff while the remaining half is carried over as surplus to the following month. Supdate(month)=Supdate(month-1)+0.5*max(0, 0.95..., , # Baseflow adjusted snowmelt runoff and rainfall runoff from TerraClimate

https://doi.org/10.5061/dryad.vx0k6dk2h

Description of the data and file structure

Input forcing of temperature, snow water equivalent, soil moisture, and precipitation comes from TerraClimate at a 1/24th degree resolution globally.

We use the runoff partitioning scheme from Qin et al., (2020) to separate runoff from snowmelt from rainfall runoff.

The original TerraClimate runoff data (q) assumed that all water (rain plus snowmelt, P) in excess of the additional soil water holding capacity (maximum soil water storage minus soil moisture or SR) and monthly evapotranspiration (ET) – or water surplus – immediately runs off that month.Â

TW =Â

S = max(0, 0.95*P - ET-SR)

q = 0.05*P + S

While this water balance approach may be valid at local scales, the approach does ...

TerraClimate dataset variables, including names, descriptions, minimum and maximum values, units, and scales.

In the study “CLIMATE-LIMITED VEGETATION CHANGE IN THE CONTERMINOUS UNITED STATES OF AMERICA”, published in the Global Change Biology journal, we evaluated the effects of climate conditions on vegetation composition and distribution in the conterminous United States (CONUS). To disentangle the direct effects of climate change from different non-climate factors, we applied "Liebig's law of the minimum" in a geospatial context, and determined the climate-limited potential for tree, shrub, herbaceous, and non-vegetation fractional cover change. We then compared these potential rates against observed change rates for the period 1986 to 2018 to identify areas of the CONUS where vegetation change is likely being limited by climatic conditions. This dataset contains the input and the resulting rasters for the study which include a) the observed rates of vegetation change, b) the climate derived potential vegetation rates of change, c) the difference between potential and observed values and d) the identified climatic limiting factor. Methods Input data

We use the available data from the “Vegetative Lifeform Cover from Landsat SR for CONUS” product (https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=1809) to evaluate the changes in vegetation fractional cover.

The information for the climate factors was derived from the TerraClimate data catalog (https://www.climatologylab.org/terraclimate.html). We downloaded data from this catalog for the period 1971 to 2018 for the following variables: minimum temperature (TMIN), precipitation (PPT), actual evapotranspiration (AET), potential evapotranspiration (PET), and climatic water deficit (DEF).

Preprocessing of vegetation fractional cover data

We resampled and aligned the maps of fractional cover using pixel averaging to the extent and resolution of the TerraClimate dataset (~ 4 km). Then, we calculated rates of lifeform cover change per pixel using the Theil-Sen slope analysis (Sen, 1968; Theil, 1992).

Preprocessing of climate variables data

To process the climate data, we defined a year time step as the months from July of one year to July of the next. Following this definition, we constructed annual maps of each climate variable for the years 1971 to 2018.

The annual maps of each climate variable were further summarized per pixel, into mean and slope (calculated as the Theil-Sen slope) across one, two, three, four, five, ten-, and 15-year lags.

Estimation of climate potential

We constructed a final multilayer dataset of response and predictor variables for the CONUS including the resulting maps of fractional cover rate of change (four response variables), the mean and slope maps for the climate variables for all the time-lags (70 predictor variables), and the initial percent cover for each lifeform in the year 1986 (four predictor variables).

We evaluated for each pixel in the CONUS which of the predictor variables produced the minimum potential rate of change in fractional cover for each lifeform class. To do that, we first calculated the 100% quantile hull of the distribution of each predictor variable against each response variable.

To calculate the 100% quantile of the predictor variables’ distribution we divided the total range of each predictor variable into equal-sized bins. The size and number of bins were set specifically per variable due to differences in their data distribution. For each of the bins, we calculated the maximum value of the vegetation rate of change, which resulted in a lookup table with the lower and upper boundaries of each bin, and the associated maximum rate of change. We constructed a total of 296 lookup tables, one per lifeform class and predictor variable combination. The resulting lookup tables were used to construct spatially explicit maps of maximum vegetation rate of change from each of the predictor variable input rasters, and the final climate potential maps were constructed by stacking all the resulting maps per lifeform class and selecting for each pixel the minimum predicted rate of change and the predictor variable that produced that rate.

Identifying climate-limited areas

We defined climate-limited areas as the parts of the CONUS with little or no differences between the estimated climate potential and the observed rates of change in fractional cover. To identify these areas, we subtracted the raster of observed rates of change from the raster of climate potential for each lifeform class.

TerraClimate 是全球陸地表面的月度氣候和氣候水資源平衡資料集。這項資料集採用氣候輔助插補法，結合 WorldClim 資料集的高空間解析度氣候常態，以及 CRU Ts4.0 和日本55 年再分析(JRA55) 的較粗空間解析度，但隨時間變化的資料。從概念上來說，此程序會套用內插的…

These data characterize active fires from MODIS, forest loss from the Global Forest Change dataset, and meteorological variables from TerraClimate. Data are summarized by 0.05 degree grid cell and by year. These data support the analyses in the following paper, which is under review in Geophysical Research Letters:Wimberly, M. C., Wanyama, D., Doughty, R., Piero, H., and Crowell, S. In Review. Increasing Fire Activity in African Tropical Forests is Associated with Land Use and Climate Change. Geophysical Research Letters.This archive contains the following files:tropical_forest_west_central_africa.shp: Ecoregion boundariestropical_forest_west_central_africa.shx: Ecoregion boundariestropical_forest_west_central_africa.dbf: Ecoregion boundariestropical_forest_west_central_africa.prj: Ecoregion boundariesAnnual_africa_modis_fd.tif: Annual count of MODIS active firesAfricaHGFCAnnualLoss2001to2021.tif: Annual count of Global Forest Change forest loss pixelsForestMask2000Prop.tif: Mean Global Forest Change forest cover in 2000ppt_mean: Annual sum of TerraClimate precipitationvpd_mean: Annual mean of TerraClimate vapor pressure deficittmax_mean: Annual mean of TerraClimate maximum temperature

We gather the precipitation data monthly scale for CHIRPS, WorldClim, and TerraClimate gridded datasets for the Upper Cauca River Basin-UCRB located to the South-West of Colombia for the 1981–2018 period. To bias-correct the precipitation data from these gridded products, we applied a methodology consisting of a point-to-pixel Quantile Mapping (QM) correction to all gridded products. The QM bias correction algorithm technique was used to correct the precipitation dataset that better reproduces the precipitation over the UCRB. The "qmap" package, developed by Lukas Gudmundsson for R software, was used for the computation. The most appropriate method (QM method) was identified as the Empirical Quantile Method, leading to selection of a best-fit precipitation series for each reference location of all gridded products. Recently, we corrected some data sets to adjust the accuracy.

Data collection

Plant hydraulic traits and height data were obtained from three sources: (1) field measurements of plant hydraulics for 210 forest species in China; (2) the TRY Plant Traits Database (https://www.try-db.org/TryWeb/Home.php; Kattge et al., 2020); and (3) published literature. For the latter we conducted searches on Web of Science, Google Scholar, and China National Knowledge Infrastructure (http://www.cnki.net) using keywords such as “hydraulic traits,” “xylem hydraulic conductivity,” “xylem vulnerability,” “water potential at 50% loss of hydraulic conductivity,” “xylem embolism resistance,” and “plant water conductivity.” A substantial portion of data in our study were obtained from published literature (Choat et al., 2012; Gleason et al., 2016) and the Xylem Functional Traits Database (XFT; https://xylemfunctionaltraits.org).

To minimize ontogenetic and methodological variation, we only included data that met the following criteria: (a) plants were grown in natural ecosystems, excluding greenhouse and common garden experiments; (b) measurements were made on adult plants and not on seedlings; (c) hydraulic traits were measured on terminal stem or branch segments in the sapwood at the crown; (d) trait data were calculated as the mean value for each species at the same site when data were from multiple sources; and (e) data values > 3 SD (standard deviation) were removed to reduce the effect of outliers (Carmona et al., 2021); (f) height data were reported at the same site where plant hydraulic traits were measured.

Climate data were obtained either from the original reports or from WorldClim version 2 (http://worldclim.org/version2; Fick & Hijmans, 2017; Table 1) if the original data were not available. The following variables measured at ~1 km² scale were extracted from WorldClim: mean annual wind speed (μ), mean annual precipitation, mean annual temperature, precipitation seasonality, temperature seasonality, wind seasonality (μS; coefficient of variation across monthly measurements × 100), precipitation of driest month, and minimum temperature of coldest month. The VPD data were extracted from the TerraClimate dataset (http://www.climatologylab.org/terraclimate.html; Abatzoglou et al., 2018). Annual PET (potential evapotranspiration) data were extracted from the CGIAR-CSI consortium (http://www.cgiar-csi.org/data; Zomer et al., 2008). Moisture index (MI), which is the ratio of precipitation to PET.

Data analysis

Trait and environment data were log₁₀-transformed to achieve approximate normality, except for P₅₀ and temperature data. We first calculated correlations among all climatic variables and for subsequent analyses retained only those variables with correlation coefficients lower than |0.7| (Dormann et al., 2013). We then ran independent multiple linear models for each trait of interest using the retained climatic variables. Model selection based on a corrected Akaike information criterion and using the R package glmulti (Calcagno & de Mazancourt, 2010), identified the best linear model for each trait. The R package ‘visreg’ (Breheny & Burchett, 2017) was used to visualize the partial relationships between wind speed and hydraulic traits. Two-dimensional contour plots were then used to explore and visualise how plant hydraulic traits varied simultaneously with wind speed and moisture index.

To quantify the strength of wind effects on plant hydraulics, models with wind parameters μ and μS included were compared to those without these wind parameters.

To test for differences in the relationship between hydraulic traits and wind speed among species grouped into different climatic regions (i.e., dry vs. wet sites, and tropical vs. temperate regions), we used standardized major axis (SMA) analyses using the R package ‘smatr’ (Warton et al., 2012). A grouping factor was added in each SMA to test whether species groups share a common slope, with p > 0.05 indicating species groups share a common slope.

Variance partitioning analysis was performed using the ‘rdacca. hp’ R package to quantify the degree to which the effect of wind speed was independent from other climatic variables (Lai et al., 2022). The individual contribution of each predictor was estimated in this analysis. This analysis also helped to illustrate the significant values of climatic variables on plant hydraulics.

A Random Forest machine-learning algorithm (implemented using the R package ‘randomForest’) was utilized to further assess the relative importance of environmental variables for each plant hydraulic trait (Breiman, 2001). To avoid multicollinearity, this analysis only included variables with correlation coefficients lower than |0.7|. A higher value of the mean decrease in accuracy (%IncMSE) indicates the increased importance of a variable (e.g., a %IncMSE value of 50 indicates that the overall mean square error would increase by 50% if that variable were to be excluded from the analysis). This provides a measure of a variable's importance in estimating the value of the target variable across the trees in the forest.

TerraClimate 是一个数据集，其中包含全球陆地表面的月度气候和气候水文平衡数据。它使用气候辅助插值，将 WorldClim 数据集中的高空间分辨率气候平均值与CRU Ts4.0 和日本55 年再分析(JRA55) 中的较低空间分辨率但随时间变化的数据相结合。从概念上讲，该程序会应用插值后的…

This data is a monthly runoff grid dataset for the China Nepal transportation corridor (50km on both sides of the China Nepal kilometer). The spatial resolution is 5km, the temporal resolution is from January 1992 to December 2022, the data format is TIFF, and the data unit is 1mm. The original data is the runoff data from the Global Land Surface Monthly Climate and Climate Water Balance Dataset released by the TerraClimate Laboratory at the University of Idaho（ https://www.climatologylab.org/terraclimate.html ）This data uses an improved Thornthwater Weather climate water balance model, which includes inputs of precipitation, evapotranspiration, soil moisture, and snowmelt, to construct publicly available TerraClimate runoff data. The original data were stored in NETCDF (.NC) format. In order to facilitate users to carry out relevant research on the transport corridor area of China and Nepal at the sub-basin scale, we used the Hydrosheds Asian 12 sub-basin vector data to perform partition statistics on the basis of the original data, and calculated the monthly total runoff in each sub-basin.

Species across the planet are shifting their ranges to track suitable climate conditions in response to climate change. Given that protected areas have higher quality habitat and often harbor higher levels of biodiversity compared to unprotected lands, it is often assumed that protected areas can serve as steppingstones for species undergoing climate-induced range shifts. However, there are several factors that may impede successful range shifts among protected areas, including the distance that must be travelled, unfavorable human land uses and climate conditions along potential movement routes, and lack of analogous climates. Through a species-agnostic lens, we evaluate these factors across the global terrestrial protected area network as measures of climate connectivity, which is defined as the ability of a landscape to facilitate or impede climate-induced movement. We found that over half of protected land areas and two-thirds of the number of protected units across the globe are at risk of climate connectivity failure, casting doubt on whether many species can successfully undergo climate-induced range shifts among protected areas. Consequently, protected areas are unlikely to serve as steppingstones for a large number of species under a warming climate. As species disappear from protected areas without commensurate immigration of species suited to the emerging climate (due to climate connectivity failure), many protected areas may be left with a depauperate suite of species under climate change. Our findings are highly relevant given recent pledges to conserve 30% of the planet by 2030 (30x30), underscore the need for innovative land management strategies that allow for species range shifts, and suggest that assisted colonization may be necessary to promote species that are adapted to the emerging climate. Methods Identifying climate analogs We followed the methods of Abatzoglou et al. (2020) and Parks et al. (2022) to characterize climate and identify backward and forward climate analogs. The specific climate variables we used were average minimum temperature of the coldest month (Tmin), average maximum temperature of the warmest month (Tmax), annual actual evapotranspiration (AET), and annual climate water deficit (CWD). AET and CWD concurrently account for evaporative demand and availability of water (N. L. Stephenson, 1990). These four variables provide complementary information pertinent to ecological systems and collectively capture the major climatic constraints on species distributions and ecological processes across a range of taxa (Dobrowski et al., 2021; Lutz et al., 2010; Parker & Abatzoglou, 2016; N. Stephenson, 1998; C. M. Williams et al., 2015). Monthly data acquired from TerraClimate (Abatzoglou et al., 2018) were used to produce these annual summaries from 1961-1990 (resolution = ~4km), which were then averaged over the same time period to represent reference period climate normals. The reference time period (1961–1990) is meant to represent climate conditions and climate niches prior to the bulk of recent warming. Future climate conditions were also computed from TerraClimate (available from www.climatologylab.org/terraclimate.html) and correspond to a 2°C increase above pre-industrial levels that are likely to manifest by mid-21st century without immediate and massive changes in global climate policies (Friedlingstein et al., 2014). As with the reference period climate, we summarized the four +2°C climate metrics annually and over a 30-year time period to represent future climate normals. All analyses in this study were conducted in the R statistical platform (R Core Team, 2020). We identified backwards and forwards analogs by estimating the climatic dissimilarity between each protected focal pixel (resolution = ~4km to match gridded climate data) and all protected pixels within a 500-km radius using a standardized Mahalanobis distance (Mahony et al., 2017). We chose the 500-km search radius as it encompasses an upper range of dispersal for some terrestrial animals and plants (Chen et al., 2011) when assuming 2°C warming by the mid-21st century; this search radius has also been used in previous studies (Bellard et al., 2014; Parks et al., 2022; J. W. Williams et al., 2007). The Mahalanobis distance metric synthesized the four climate variables (i.e. Tmin, Tmax, AET, and CWD; fig. 2a) by measuring distance in multivariate space away from a centroid using principal components analysis of standardized anomalies. Mahalanobis distance scales multivariate mean climate conditions between a pixel and those within the search radius by the focal pixel’s covariance and magnitude of interannual climate variability (ICV) across the four metrics. For backwards analogs, we characterized +2°C ICV and reference period climate normals to calculate climatic dissimilarity; for forward analogs, we used reference period ICV and +2°C climatic normals to calculate climatic dissimilarity. We standardized Mahalanobis distance to account for data dimensionality by calculating a multivariate z-score (σd) based on a Chi distribution (Mahony et al., 2017). σd represents the climate similarity between each focal pixel and its candidate backward and forward analogs (i.e. all other protected terrestrial pixels within 500 km), and we considered any protected pixels with σd ≤ 0.5 as climate analogs (fig. 2b) (following Parks et al., 2022). We were unable to calculate Mahalanobis distance when there was no ICV for any one of the four variables, and as a consequence, these areas are omitted from all analyses; this affects, for example, a relatively small tropical area in South America (CWD=0 each year) and areas perennially covered by snow (CWD=0 each year; e.g. most of Greenland). We focused our analyses on protected areas as defined by the World Database on Protected Areas (WDPA) (IUCN & UNEP-WCMC, 2019) and included protected areas classified as IUCN (International Union of Conservation for Nature) Management Categories I-VI, except those identified as ‘proposed’, ‘marine’, or otherwise aquatic (e.g. wetland, riverine, endorheic). A large number of protected areas, however, were not assigned an IUCN category in the WDPA (identified as ‘Not Reported’, ‘Not Assigned’, or ‘Not Applicable’) but are likely to have reasonably high levels of protection (e.g. Kruger National Park in South Africa). We included these additional protected areas if the level of human modification was similar or less than that observed within IUCN category I-VI protected areas. To do so, we measured mean land-use intensity within each IUCN category I-VI protected area using the Human Modification Gradient (HMG) raster dataset (Kennedy et al., 2019) and calculated the 80th percentile of the resulting distribution. Any unassigned protected areas with a mean HMG less than or equal to this identified threshold were included in our study (following Dobrowski et al., 2021). We then converted this vector-based polygon dataset to raster format (resolution = ~4km to match gridded climate data; n=1,063,748 pixels). It is well-recognized that the WDPA contains a large number of duplicate and overlapping polygons (Palfrey et al., 2022; Vimal et al., 2021). Although this does not affect summaries across the globe or for individual countries (described below), it provides a challenge when trying to summarize by individual protected areas (due to double-counting). Consequently, we ‘cleaned’ the WDPA prior to summarizing the climate connectivity metrics for individual protected areas by removing polygons that exhibited ≥ 90% overlap with another; this resulted in 29,752 individual protected areas (available in the Electronic Supplemental Material). Least-cost path modelling Following Dobrowski and Parks (2016) and Carroll et al. (2018), we used least-cost path modelling (Adriaensen et al. 2003) to build potential climate-induced movement routes between each protected focal pixel and its backward and forward analogs. The least-cost models were parameterized with resistance surfaces based on climate dissimilarity and the human modification gradient (HMG) (Kennedy et al., 2019). For backward analog modelling, we characterized climatic dissimilarity (i.e. climatic resistance) using two intermediate surfaces, the first being the Mahalanobis distance between each focal pixel (using +2°C ICV) and all other pixels using reference period climate normals (fig. 2c) and the second being the Mahalanobis distance (using +2°C ICV) and all other pixels using +2°C climate normals (fig. 2d). These two surfaces provide a proxy for climate similarity designed to capture transient changes between the reference period and +2°C climate; these were then averaged to characterize the overall climatic resistance across time and space (fig. 2d). For forward analog modelling, the process is similar except we used reference period ICV when characterizing climatic resistance (fig. 2a-2d). We then multiplied the climatic resistance (fig. 2d) by HMG (fig. 2e) to create the final resistance surface for least-cost path modeling (cf. Parks et al., 2020). Prior to this step, we rescaled HMG from its native range (0–1) to 1–25 to correspond with the range of Mahalanobis distance values and thereby grant comparable weights to climatic resistance and HMG resistance (~95% of all Mahalanobis distance values are below 25 within a 500km radius). Open water was given a resistance=25 so that paths would avoid water when possible. Least-cost path modelling was achieved using the gdistance package (van Etten, 2017); paths represent the least accumulated cost across the final resistance surface (fig. 2f) between each focal pixel and analog (fig. 2g). Because paths were rarely straight lines, some were longer than the 500km that we established as a search radius. We removed these longer paths to abide by the biologically informed upper dispersal

TerraClimate là một tập dữ liệu về khí hậu và cân bằng nước khí hậu hằng tháng cho các bề mặt trên đất liền trên toàn cầu. Nền tảng này sử dụng phương pháp nội suy có sự hỗ trợ của khí hậu, kết hợp các giá trị trung bình khí hậu có độ phân giải không gian cao từ tập dữ liệu WorldClim, với độ phân giải không gian thô hơn, nhưng dữ liệu thay đổi theo thời gian từ CRU Ts4.0 và Phân tích lại 55 năm của Nhật Bản (JRA55). Về mặt khái niệm, quy trình này áp dụng các điểm bất thường thay đổi theo thời gian được nội suy từ CRU Ts4.0/JRA55 vào khí hậu học có độ phân giải không gian cao của WorldClim để tạo một tập dữ liệu có độ phân giải không gian cao bao gồm một bản ghi thời gian rộng hơn. Thông tin về thời gian được kế thừa từ CRU Ts4.0 cho hầu hết các bề mặt đất trên toàn cầu về nhiệt độ, lượng mưa và áp suất hơi. Tuy nhiên, dữ liệu JRA55 được dùng cho những khu vực mà dữ liệu CRU không có trạm khí hậu nào đóng góp (bao gồm toàn bộ Nam Cực và một số khu vực ở Châu Phi, Nam Mỹ và các hòn đảo rải rác). Đối với các biến khí hậu chính về nhiệt độ, áp suất hơi và lượng mưa, Đại học Idaho cung cấp thêm dữ liệu về số lượng trạm (từ 0 đến 8) đã đóng góp vào dữ liệu CRU Ts4.0 mà TerraClimate sử dụng. JRA55 chỉ được dùng cho bức xạ mặt trời và tốc độ gió. Ngoài ra, TerraClimate còn tạo ra các tập dữ liệu cân bằng nước bề mặt hằng tháng bằng cách sử dụng mô hình cân bằng nước kết hợp lượng bốc hơi tham chiếu, lượng mưa, nhiệt độ và khả năng giữ nước có thể chiết xuất của cây trồng được nội suy. Mô hình cân bằng nước khí hậu Thornthwaite-Mather đã được sửa đổi và dữ liệu về khả năng lưu trữ nước có thể chiết xuất trong đất được sử dụng ở lưới 0,5° của Wang-Erlandsson và cộng sự (2016). Hạn chế về dữ liệu: Xu hướng dài hạn trong dữ liệu được kế thừa từ các tập dữ liệu gốc. Bạn không nên sử dụng TerraClimate trực tiếp để đánh giá độc lập các xu hướng. TerraClimate sẽ không ghi lại sự biến đổi theo thời gian ở quy mô nhỏ hơn so với các tập dữ liệu mẹ và do đó không thể ghi lại sự biến đổi về tỷ lệ và sự đảo ngược lượng mưa do địa hình. Mô hình cân bằng nước rất đơn giản và không tính đến sự không đồng nhất về các loại thảm thực vật hoặc phản ứng sinh lý của chúng đối với các điều kiện môi trường thay đổi. Xác thực có giới hạn ở những khu vực có ít dữ liệu (ví dụ: Nam Cực).

This archive contains a dataset of high-spatial resolution (1/24 , ~4-km) monthly climate and climatic water balance for global terrestrial surfaces from 1958-2015. These data were created by using climatically aided interpolation, combining high-spatial resolution climatological normals from the WorldClim version 1.4 and version 2 datasets, with coarser resolution time varying (i.e. monthly) data from CRU Ts4.0 and JRA-55 to produce a monthly dataset of precipitation, maximum and minimum temperature, wind speed, vapor pressure, and solar radiation. TerraClimate additionally produces monthly surface water balance datasets using a water balance model that incorporates reference evapotranspiration, precipitation, temperature, and interpolated plant extractable soil water capacity.