Search for a command to run...
These data were compiled to understand the effects of land use and riparian vegetation composition, structure and health on local abundance and species diversity of birds along five reaches of the Colorado River delta in Mexico that have experienced restoration through active revegetation and environmental water deliveries. The data release is composed of one .csv file “BirdVegLandCRD2002to2021_Data20251010.csv” that contains the data and one .xlsx file “BirdVegLandCRD2002to2021_Metadata20251010.xlsx” with metadata (i.e., explanation of data columns). We provided a summary of data collection methods in the following paragraphs. The dataset includes bird species detections, cover of riparian vegetation by life form and species, vegetation reflectance indices, estimates of evapotranspiration, and the spatial configuration of seven land cover classes along 27 routes comprised of 6 to 20 bird count stations (total of 230 bird count stations), which were distributed across the five river reaches within both active restoration areas (5 sites, 7 routes) and control (unrestored) sites (20 routes). The number of stations along restored sites varied as a function of their size, and new routes were established as soon as restoration efforts were completed (sites were restored in Reach 2 and 4 only, from 2010 to 2017). Control routes typically included 8 bird point count stations spaced a minimum of 200 m apart, for approximately 2 km along the upstream-downstream river axis and were located within the floodplain portion not occupied by human activities and not actively restored (although we cannot exclude some effects of environmental flows). Observations of birds and vegetation correspond to the period 2002 to 2021, while land use was analyzed for the year 2019 only. At each bird count station, we conducted three surveys (May 15-31, June 15-30, and July 15-31) yearly once we established each route. We designed the survey to detect multiple species of all-year and summer resident birds (“breeding birds”), using three survey periods to detect species with different breeding phenology. The period May-July includes the peak of the reproductive period for most breeding species in the CRD. This strategy differs from one that treats multiple breeding-season surveys (or multiple visits within the same year) as replicate counts for any particular species. We followed a variable distance point count protocol (Reynolds et al. 1980; Rosenstock et al. 2002) in which teams of at least two observers visited point-count stations on sunny days with calm winds (<5 km per hour), from 30 minutes before sunrise to a maximum of four hours after sunrise. Observers recorded all bird species and the number of individuals heard or seen during a five-minute period at each station. We estimated the distance between the observer and the location of individual detections to fall within three predetermined fixed-radii circles (0-25, 26-50, 51-200 m) around the observer. We did not record birds observed flying over but not using the habitat. We used concurrent vocalizations and songs to determine if multiple individuals were present and to avoid double-counting of individuals. The final database included 163 bird taxa and 103,358 individual detections. Only one taxon could not be identified to the species level. We conducted on-the-ground vegetation surveys in 2002, 2007, 2009, 2010 and every year from 2013 to 2021 in all bird point count stations. From 2002 to 2010, we only sampled the control sites in Reaches 3, 4, and 5. We initiated vegetation sampling in the control sites in Reach 1 in 2016. Vegetation surveys consisted of visual estimation of the percent cover (to the nearest 1% on a scale from 0 to 100%) of trees (plants taller than 3 m), shrubs (0.5 and 3 m), tall herbs (0.1 - 0.5 m), low herbs (<0.1 m), marshland vegetation, as well as bare soil and open water within a 50 m radius from the center of the bird point count station. We recorded cover values in each category as relative cover; thus, their sum was 100%. For the tree, shrub, tall herb, and low herb vegetation strata, we also estimated the percent cover of each individual species within each stratum using the same relative cover scale (i.e., sum of 100% within each stratum). Because we recorded species cover for each vegetation stratum; the same species could contribute to relative cover in more than one stratum. We handled marshland vegetation differently. We did not note cover values by height class: we only recorded a single relative cover value for species defined as characteristic of marshlands (reeds, cattails, sedges, and rushes). We did not record aquatic plants. For each plant species we calculated an aggregate cover for each observation (bird count station sampled at a given year) as the mean cover of the species across all strata, using the relative cover of each stratum (including bare soil) as weighting factor in the species mean calculation. Vegetation reflectance indices and evapotranspiration (ET) data were calculated at varying spatial scales (100 m, 500 m, 1 km, and 2 km from the center of bird count stations). Vegetation reflectance indices (VIs): the Normalized Difference Vegetation Index (NDVI), the Enhanced Vegetation Index (EVI), and the two-band EVI (EVI2) were obtained from Landsat imagery with a biweekly temporal frequency and covering the entire period of bird surveys (2002-2021). NDVI, EVI, and EVI2 were produced as tabular data sets. ET was then produced using EVI and EVI2. Also, ET was calculated from EVI2 using an empirical equation (Nagler et al., 2013). Data acquisition and processing were performed at the University of Arizona VIP lab (vip.arizona.edu) using their large Linux cluster of computing and storage resources. We weighted (averaged) the VI and ET values by the pixel area within the buffer by transforming the raster cells to polygons in ArcGIS Pro. We computed mean values using the 15 May – 31 July period for each one year at each bird count station, thus matching the observations from the bird surveys. We only released these mean values. Biweekly minimum, maximum, average and standard deviation from all pixels within the bird count station, and more methodological details for data collection and processing are available at the associated data release: Nagler et al. (2022). Note that Nagler et al. (2022) released values from 2000 to 2020 and we used the 2002-2021 period for mean calculations. The 2021 data were obtained from Nagler’s lab but, by the time of this data release, remain unpublished. We used a land cover map based on 2019 imagery created by Gómez-Sapiens (2025) to determine the spatial configuration of land cover classes around the bird count stations with a battery of landscape characteristics using the landscapemetrics package (Hesselbarth et al. 2019) in R version 3.5.1 (R Core Team, 2020) – see metadata for list of landscape characteristics. The map has seven land cover classes: “water”, “barren”, “shrubland”, “riparian forest”, “marshland”, “agriculture”, and “urban areas”. Unless otherwise specified, all were calculated at buffer distances of 100 m, 500 m, 1 km, 2 km, and 5 km radius from the center of the bird count station, for all land cover types combined, and for each of the seven land cover types. The data released here have been used in the following four publications: Grand et al. (2024), and González-Sargas et al. (2024a; 2024b; 2026). We refer the user of this data to these publications for in-depth descriptions on the study system, including the biophysical and historical context for restoration, the study design, data collection methods, analyses, and interpretations. References cited: Grand J, Meehan TD, DeLuca WV, Morton J, Pitt J, Calvo-Fonseca A, Dodge C, Gómez-Sapiens M, González-Sargas E, Hinojosa-Huerta O, Nagler P, Restrepo-Giraldo C, Shafroth PB, Villagomez-Palma S, Wilsey CB. (2024) Strategic restoration planning for land birds in the Colorado River delta, Mexico. Journal of Environmental Management 351:119755. https://doi.org/10.1016/j.jenvman.2023.119755 Gómez-Sapiens M (2025). 2019 Land cover map of riparian corridor and adjacent areas in the Colorado River Delta, Mexico. Zenodo. https://doi.org/10.5281/zenodo.14816607 González-Sargas E, Gómez-Sapiens M, Hinojosa-Huerta O, Villagomez-Palma S, Calvo-Fonseca A, Grand J, Meehan TD, Dodge C, Nagler PL, Restrepo-Giraldo C, Nieblas C, Meléndez A, Real Rangel R, Shafroth PB. (2024a) Avian communities respond to plant and landscape composition in actively revegetated floodplains of the Colorado River delta in Mexico. Ecological Engineering 205:107266. https://doi.org/10.1016/j.ecoleng.2024.107266 González-Sargas E, Meehan TD, Hinojosa-Huerta O, Villagómez-Palma S, Calvo-Fonseca A, Dodge C, Gómez-Sapiens M, Shafroth PB. (2024b) Bird community response to one decade of riparian restoration along the Colorado River Delta in Mexico. Ecological Engineering 205:107291. https://doi.org/10.1016/j.ecoleng.2024.107291 González-Sargas E, Meehan TD, Hinojosa-Huerta O, Villagomez-Palma S, Dodge C, Gómez-Sapiens M, Nagler PL, Shafroth PB. (2026) Bird guilds exhibit varied responses to floodplain forest restoration in the Colorado River delta, Mexico. Journal of Arid Environments 234:105558. https://doi.org/10.1016/j.jaridenv.2026.105558 McGarigal K, Marks BJ (1995) FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. Gen. Tech. Rep. PNW-GTR-351. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 122 p Nagler PL, Glenn EP, Nguyen U, Scott R, Doody T (2013) Estimating riparian and agricultural actual evapotranspiration by reference evapotranspiration and MODIS Enhanced Vegetation Index. Remote Sensing 5:3849-3871. https://doi.org/10.3390/rs5083849 Nagler PL, Barreto-Muñoz A, Didan K, González E, Shafroth PB, Gómez-Sapiens M (2022) Colorado River Delta Project: