The timing of life history events around reproduction and early development is critical in population dynamics, and it can determine recruitment success, species dispersal, and population connectivity. In ectotherms, as well as in plants and fungi, phenology is mediated by the nonlinear effects of temperature on physiology and development, meaning that spatiotemporal variation in temperature can exert powerful controls on the timing of local reproduction and recruitment. Here, we examine reproduction phenology (fertilization of embryos, duration of embryonic development during brooding, and larval release) of the intertidal acorn barnacle Semibalanus balanoides in 2002-04 and in 2019-24 at up to 8 sites along a steep temperature gradient in the northwest Atlantic Ocean. At each site and year, we assessed how phenology varied with intertidal temperature, estimated with a hybrid atmosphere-ocean data assimilation model. Although within-site reproduction was delayed due to interannual and decadal fall warming (3.7 days per 1°C), fertilization at all sites in all years still occurred within a 1-month timeframe. In contrast, latitudinal differences in intertidal temperature resulted in substantially different brooding durations (up to 95 days difference) and, by extension, larval release timing (e.g., Dec 18 vs Apr 4). Consequently, lower latitude larvae tended to enter the water column at the start of winter, while higher latitude larvae were not released until spring. These different larval release times result in regional differences in temperature-mediated larval development, potentially resulting in lower latitude populations experiencing greater dispersal. Our study is one of the first to evaluate these relationships through both space and time in natural populations, and we show that both spatial gradients and interannual variation in the seasonal temperature cycle can mediate reproductive physiology and dispersal of temperate and polar species.

The data package contains two folders. The first folder data/ contains: (i) weekly barnacle reproductive proportion data, (ii) hourly temperature logger data from Newagen, Maine, 2003, (iii) filtered intertidal temperatures, and (iv) site metadata. The second folder NOAH_IntertidalModel/ contains: (v) North American Regional Reanalysis (NARR) example data for the year 2000, (vi) OSTIA reprocessed SST example data for the year 2000, subset for Northeast US/Canada coast, and (vii) the suggested directory structure for using XTide, along with the header file for the version of libtcd (a prerequisite software for XTide) used in this study.

(i) Weekly barnacle reproductive proportions were measured at up to eight sites along a steep temperature gradient in the northwest Atlantic during two historical and five modern years.

(ii) We include hourly temperature logger (Onset Corp.) data for Newagen, Maine, collected in 2003 for intertidal and subtidal habitat, as well as logger data for Halifax, Nova Scotia, Canada (2003), and Falmouth, Massachusetts (2004 and 2022).

(iii) We also include filtered, hourly intertidal temperature estimates for all sites and years included in this study. These were generated via a hybrid atmosphere-ocean data assimilation model (using the Noah land surface temperature model v1.91, modified to simulate intertidal temperature and parameterized for Semibalanus balanoides body temperature). The modified Noah Intertidal Temperature model simulates heat transfer between the atmosphere, a biotic layer of S. balanoides barnacles, and a substrate of solid granite, with periodic tidal flooding estimated with the XTide tide prediction software (v2.8.2, https://flaterco.com/xtide/xtide.html). For atmospheric forcing, we used the National Centers for Environmental Prediction (NCEP) North American Regional Reanalysis (NARR) data products (see below). For ocean temperature forcing, we used the Operational Sea Surface Temperature and Sea Ice Analysis (OSTIA) reprocessed and near-real-time data products (see below). The Noah Land Surface Temperature model software is available for download from https://ral.ucar.edu/model/unified-noah-lsm. Modeled temperatures were filtered with the low-pass PL64 filter using a cutoff of 33 hours to remove high-frequency variability, yielding the filtered temperatures included here. We also include filtered versions of the temperature logger measurements. 

(iv) The file site_locations.csv includes metadata for the study, including site names, locations (lat/lon), and parameters used in the intertidal temperature model.

(v) NARR data were downloaded from https://downloads.psl.noaa.gov/Datasets/NARR/monolevel/. We include example data for the year 2000.

(vi) OSTIA reprocessed data were obtained from https://data.marine.copernicus.eu/product/SST_GLO_SST_L4_REP_OBSERVATIONS_010_011/description. We include example data for the year 2000. OSTIA near real-time data were also used in this study and can be obtained from https://data.marine.copernicus.eu/product/SST_GLO_SST_L4_NRT_OBSERVATIONS_010_001/description 

(vii) Here. We include the recommended directory structure for using the XTide software, along with the header file (nerr.tcd) for the libtcd v2.2.4 software used in this study. Both XTide and libtcd can be downloaded here: https://flaterco.com/xtide/files.html

Harmonics data, for estimating tidal height via XTide software, were downloaded from https://flaterco.com/xtide/files.html#harmonicsfiles. Note that current versions of these files, including those used in this study, no longer extend outside of US waters, though data can be downloaded directly from individual data sources. See why here: https://flaterco.com/xtide/faq.html#60. The harmonic constants "used to perform tide predictions for locations in Canada are derived from sea level data made available by Marine Environmental Data Services, Fisheries and Oceans Canada, for public, non-commercial use." Therefore, these files can be made available for non-commercial use upon request.