Skip to main content
Dryad

Positive correlation between Ammospiza caudacuta (Saltmarsh Sparrow) capture and productivity supports use of a novel rapid assessment monitoring protocol

Cite this dataset

Sanchez Jr., Arthur et al. (2024). Positive correlation between Ammospiza caudacuta (Saltmarsh Sparrow) capture and productivity supports use of a novel rapid assessment monitoring protocol [Dataset]. Dryad. https://doi.org/10.5061/dryad.q573n5trr

Abstract

Salt marshes in the northeastern United States provide critical breeding habitat for tidal marsh specialist birds like the Saltmarsh Sparrow (Ammospiza caudacuta). The Saltmarsh Sparrow population has declined by nine percent annually, necessitating immediate conservation actions for this vulnerable species. However, estimating species vital rates across a large geographic region is logistically challenging and cost prohibitive. Therefore, we developed and tested a rapid assessment monitoring protocol focused on reproductive metrics to enhance future conservation planning.

We used three years of intensive demographic data from 12 sites to estimate daily nest survival, nest period success, fledglings produced per female, and successful broods per female. We implemented the rapid assessment protocol co-located at intensive sites in the same years to estimate the number of captured females and juveniles. We used Pearson’s correlation analyses to determine the association of intensive metrics with rapid metrics. We found that the sum of Saltmarsh Sparrow female and juvenile captures was positively correlated with daily nest survival, nest period success, fledglings produced per female, and successful broods per female. Our results demonstrate that fixed-effort mist-netting from our rapid assessment protocol is an informative and time-efficient sampling method that can aid in making informed management decisions related to Saltmarsh Sparrow conservation.

README: Positive correlation between Ammospiza caudacuta (Saltmarsh Sparrow) capture and productivity supports use of a novel rapid assessment monitoring protocol

https://doi.org/10.5061/dryad.q573n5trr

The dataset contents include mark-recapture data, nest monitoring data, and a file that includes productivity data from nest survival analyses (e.g., daily nest survival, nest period success, fledglings produced per female, and broods produced per female) and raw capture data of females and juveniles. All data was collected in tidal marshes throughout the global breeding range of the Ammospiza caudacuta (Saltmarsh Sparrow). 

Description of the data and file structure

All of the files here are Microsoft excel worksheets and have been proofed and processed.

MCestimateSALS_RDcensored.xls - This file includes all of the nest monitoring data. By column, we have SHARPNestID which identifies each nest and the nests are numbered in chronological order. For example, the first nest is AT21002. You can interpret this as AT was the site at which the nest was found, 21 is the year of monitoring (we used the last two digits of the year; 2021), and it was the second (002) nest that was found (AT21002). We have 12 study areas in the Site column. The sites are identified as follows: AT = AT&T (in New Jersey), BI = Barn Island (in Connecticut), CL = Chapman's Landing (in New Hampshire), EL = Eldridge (in Maine), FB =Furbish Marsh (in Maine), HM = Hammonasset (in Connecticut), ID =Idlewild (in New York), MN = Marine Nature Preserve (in New York), NC = North Cinder Island (in New York), OC = Oyster Creek (in New Jersey), SG = Saxis and Guard (in Virginia), and Wallops Island (in = Virginia). Furthermore, Chapman's Landing was sampled twice, once in 2019 and again in 2021. Idlewild was sampled twice, once in 2018 and again in 2019. Oyster Creek was also sampled twice, once in 2019 and again in 2021. Lastly, Hammonasset was sampled three times, once in 2018, and again in 2019 and 2021. We treated each year and site combination individually totaling in 12 sites, where 4 sites were sampled for multiple years, resulting in 17 unique site+year combinations. Year, Species, Easting, Northing, and FemaleBandNum (Band Number) is self-explanatory. UltimateNestFate is a category assigned to each nest that was monitored. V1-V11 are nest visits with each column representing an individual visit. This file is directly set up to be imported into MCestimate which is a program that uses Markov Chain algorithms to estimate daily nest survival probabilities. MCestimate needs a minimum of two visits to estimate nest failure and this file has already been censored following Etterson et al. (2007) and Stanley (2004). You can use some of the outputs in MCNest to estimate seasonal fecundity but there are other parameters required for MCNest. In our supplemental materials, we have an example table of all of the parameters that need to be imported to MCNest in order to estimate measure of fecundity.

In this file, there are 5 type errors. The type errors are dates that did not read in correctly when uploading the data to Dryad. The dates have been double checked and are good to be uploading into MCestimate for analysis.

mo_state_f_survival.xls - This file includes all of the mark-recapture data we used to adjust rapid demographic captures. The mark-recapture data was used from 12 intensive demographic sites from 2011-2021. This file includes band number, species, age (AHY = After Hatch Year), sex (BP = Brood Patch; Female), site (same naming convention as the file above), state, year, latitude, interval1-4 (int1, int2, etc.), character (ch), species, and state. The intervals were set in month during the breeding season (May, June, July, and August). We used this data in an open-population Cormack Jolly-Seber Survival model in program R using the RMark package to estimate capture probabilities during the breeding season (May, June, July, and August). The columns you need to execute this analysis is character, species, and state (column O, P, and Q). These data will need to be converted to a txt. file to be imported in to R and to create the capture matrix in RMark. We used time (which are structured as 1 month intervals during the breeding season), species, and state in a global model to estimate capture probabilities in each state. We then divided the raw number of captures by the appropriate capture probability depending on where they were captured geographically (State and Site).

In this file, there are 4 blank labels. The blank label columns are empty and contain no data. There are 3 duplicate labels. The duplicate labels are species, state, and year, but as mentioned previously, these data will need to be converted to a txt. file to import into R and for RMark to create a capture matrix. To do so, you will only need columns O, P, and Q. Lastly, there are 3 type errors which are in row 2916 through 2932. These rows do not have a dash in the 9-digit band number. For example, the band number in row 2916 should read 2811-38210 but reads as 281138210. The band numbers were used to identify individuals but are not directly used in the open-population Cormack Jolly-Seber Survival model.

Regression_Parms_SALScensoredV2.xls - This file is an all-consuming file. This file includes Site (intensive demographic site), Year, nest survival outputs (Daily nest survival, nest period success), fecundity outputs (fledglings produced per female; Fledglings, broods produced per female; Broods), RD_Site (Rapid assessment site; paired with intensive sites in column A), state, RD_V1Date (RD 1st visit), RD_V2Date (RD 2nd visit), CapProbRD_V1 (Capture Probability for first RD visit), CapProbRD_V2 (Capture probability for second RD visit visit), RD_V1FemaleCaps (females captured in 1st RD visit), RD_V1JuvCaps (juveniles captured in 1st RD visit), RD_V2FemaleCaps (females captured in 2nd RD visit), RD_V2JuvCaps (juveniles captured in 2nd RD visit), RD_FemaleCaps (total number of females captured in both RD visits), RD_JuvCaps (total number of juveniles captured in both RD visits), AdjCountFemaleV1 (adjusted females captured in RD visit one adjusted by CapProbRD_V1), AdjCountJuvV1 (adjusted juveniles captured in RD visit one adjusted by CapProbRD_V1), AdjCountFemaleV2 (adjusted females captured in RD visit two adjusted by CapProbRD_V2), AdjCountJuvV2 (adjusted juveniles captured in RD visit two adjusted by CapProbRD_V2), AdjFemale (total number of adjusted females captured), AdjJuv (total number of adjusted juveniles captured), AdjFemaleJuv (sum of total number of adjusted females and juveniles captured).

The aforementioned file is a lot to take in but the next file is more straight forward and derived from the all-consuming file. For now, know that we used the Regression_Parms_SALScensoredV2.xls file to run regression and correlation models.

SCC_Raw_SALS - This file is very straight forward and can also be used for regression and correlation models. This file incudes Daily nest survival, nest period success, fecundity outputs (fledglings produced per female; Fledglings, broods produced per female; Broods), and the unadjusted female and juvenile captures. RD_FemaleCaps (Total females captured), RD_JuvCaps (total juveniles captured), RD_FemalesJuvs (the sum of total females and juveniles captured). All of the data in this file has been normalized.

Methods

Our data was collected through numerous field seasons by trained surveyors in tidal mashes across the northeastern United States. At each field site an array of different survey methods were utilized inculding mark-recpature techniques for survival analyses, and nest searching and monitoring for nest survival analyses. Our survey data was imported to the SHARP Demographic Database (https://sharp-qa.azurewebsites.net/) and maintained and QAQC'd by researchers at collaborating universities (University of Maine, University of Nuw Hampshire, University of Connecticut, State UNiversity of New York SChool of Environmental Science and Forestry, and University of Delaware).

 

Funding

National Fish and Wildlife Foundation, Award: F18AC0046

United States Fish and Wildlife Service, Award: F21AC00004

Natural Resources Conservation Service