Environmental heterogeneity plays a bigger role than diet quality in driving divergent California sea lion population trends

1 1 2 3 Environmental heterogeneity plays a bigger role than diet 4 quality in driving divergent California sea lion population 5 trends 6 7 Diet quality and population trends of sea lions 8 9 10 11 12 Ana Lucía Pozas-Franco1, *, David. A. S. Rosen1,¶, Andrew W. Trites1, ¶, Francisco J. García- 13 Rodríguez2, Claudia J. Hernández-Camacho2, ¶ 14 15 16 1Marine Mammal Research Unit, Institute for the Oceans and Fisheries, University of British Columbia, Vancouver 17 BC, V6T 1Z4, Canada 18 19 2Instituto Politécnico Nacional, Centro Interdisciplinario de Ciencias Marinas, La Paz, Baja California Sur, 23096, 20 México 21 22 23 24 *Corresponding author 25 E-mail: [email protected] 26 27 ¶These authors contributed equally to this work and were thesis committee members. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 2 28 Abstract 29 30 While the global population of California sea lions (Zalophus californianus) is increasing, regional 31 trends show a decline in the Gulf of California (GoC, Mexico) and an increase in the Channel 32 Islands (CI; California, U.S.) over the last 40 years. The drivers of these divergent trends remain 33 unclear, but previous pinniped studies suggest that differences in diet quality—rather than prey 34 abundance—may play a role. We therefore conducted an analysis to examine how sea lion 35 population trajectories relate to diet quality, specifically looking at diet energy density and diet 36 diversity. Using population and diet data from 1980 to 2020 for sea lions in the GoC and CI, we 37 found no simple relationships between population trajectories and diet quality over time. Energy 38 densities of sea lion diets were similar between the two regions, but GoC sea lions consumed a 39 more diverse range of prey (n = 88 vs. 23 main prey species) dominated by benthic species and 40 schooling fishes, while CI diets consisted mainly of schooling fishes and squid. We also found that 41 GoC sea lions ate more benthic prey and less schooling fish during the 2014–2016 heatwave— 42 decreasing their overall diet energy density, similar to the CI. This shift coincided with a temporary 43 population decline in the CI but had variable effects on GoC populations. Overall, our findings 44 suggest that regional population trends are influenced by complex ecological factors beyond diet 45 quality alone, highlighting the need to consider environmental variability and prey composition 46 when assessing the resilience of sea lion populations to climate-driven changes. 47 48 49 KEYWORDS: Diet quality, Population decline, Zalophus californianus, Gulf of California, 50 Channel Islands, Pinniped, Environmental heterogeneity .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 3 51 1. Introduction 52 53 California sea lions (Zalophus californianus ) are widely distributed along the Pacific coast of 54 North America from British Columbia, Canada to the Gulf of California, Mexico, but only 55 reproduce on certain islands (rookeries) along the southern coast of California (U.S.), the Mexican 56 Pacific, and the Gulf of California (1,2). Most of the global population (80%) breeds in California 57 where numbers increased at an annual rate of 2.9% between 1964–2014 (3). The remainder of the 58 population (20%) breeds in Mexico, where—with a few exceptions—most rookeries show a 59 declining trend. In the Gulf of California, sea lion populations experienced on average a 2% decline 60 per year between 1984–2015 (4,5). 61 62 In the southern Pacific coast of the U.S., California sea lions breed almost exclusively at four 63 rookeries that form part of the Channel Islands which vary in size from 3,000–60,000 individuals 64 (6). This population has grown steadily since the 1980’s but has experienced temporary population 65 declines in some years associated with increased sea surface temperatures as seen during the 2012– 66 2016 marine heatwave (6). The population has since recovered, totaling 111,713 sea lions in 2019 67 (6). In contrast, breeding sea lion populations in the Gulf of California, Mexico, are distributed 68 among 13 rookeries of varying size from 400–6,000 individuals (1,2). Although these rookeries 69 vary in population growth, most show a declining trajectory since the 1980’s. Only one rookery in 70 the southernmost Gulf of California, Los Islotes, is considered to have a population that has been 71 increasing since 1979 (7). 72 73 The underlying factors causing a divergence in California sea lion population trajectories in the 74 Channel Islands compared to the Gulf of California are still unknown (5). Possible contributing .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 4 75 factors that are known to affect marine mammal species in general include regional differences in 76 prey availability, pollution (both chemical and noise), disease, biotoxins, fishing gear 77 entanglements, anthropogenic mortality (disturbance, legal and illegal shooting), and migration 78 (8–11). Of these contributing factors, regional differences in diets associated with environmental 79 change have been identified as the most likely contributor to population trajectories in several 80 species (4). 81 82 Previous studies have generally focused on the negative effects that short-term reductions in prey 83 abundance have on California sea lion numbers at rookeries in the Channel Islands (3,12,13) and 84 the Gulf of California (7,14–18). Sharp declines in quantities of primary prey species available to 85 sea lions are known to occur during El Niño events in California when warm water causes prey to 86 remain at inaccessible depths, leading to increased pup mortality (19). However, El Niño events 87 do not appear to have a comparable direct effect on the Gulf of California sea lion populations 88 where the response to warming events depends on location, population size, and regional dynamics 89 (5). This leads to the question of whether changes in the quality of prey (rather that changes in 90 quantity) might better explain differences that have occurred in sea lion numbers over a longer 91 timeframe (20), as suggested for Steller sea lions (21–25). 92 93 Diet quality can be assessed in many ways. Two main metrics of diet quality are diet energy density 94 (an important aspect of the nutritional value of prey species) and diet diversity (i.e., the variety of 95 species that compose the diet). These diet characteristics can affect the nutritional status of 96 individuals, their reproduction and survival rates, as well as their susceptibility to disease and 97 predation (23,24). Based on broad trends observed among different marine mammals, a diet .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 5 98 dominated by a few energy-rich species would be hypothesized to support a growing population, 99 while a switch to a more diverse diet of energy-poor species would be expected to cause population 100 declines (26). It has also been demonstrated that changes in environmental conditions can alter the 101 diet composition and ‘quality’ of prey species available to predators. However, it is not known 102 how such differences or changes in diet quality may be influencing the population dynamics of 103 California sea lions in the Channel Islands compared to the Gulf of California. 104 To investigate the effect of diet on population trajectories, we used estimates of average diet energy 105 density, and two measures of diet diversity to quantify diet quality of California sea lions at the 106 Channel Islands and the Gulf of California rookeries from 1980–2020. We compared the measures 107 of diet quality between and within the two geographic regions and tested for relationships between 108 rates of population change and the different measures of diet quality over time. Finally, the effects 109 of increased sea surface temperatures (2014–2016) on sea lion diet and populations were also 110 compared between regions. Obtaining a better understanding of the interplay between 111 environmental changes, diets, and population trajectories is needed to inform policies regarding 112 the conservation and management of California sea lions in Mexico and the U.S. 113 114 2. Materials and Methods 115 116 2.1 Population and diet data 117 This meta-analysis focused on the four California sea lion rookeries in the Channel Islands (San 118 Miguel, San Clemente, Santa Barbara, and San Nicolas), and 12 of the 13 rookeries along the Gulf .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 6 119 of California (Fig.1). We omitted San Jorge due to a lack of available data and also had insufficient 120 data to include the west coast of the Baja California Peninsula. Available population and diet data 121 for California sea lion rookeries in the Channel Islands and the Gulf of California from 1980–2020 122 were gathered from published and unpublished sources (Table S1). All data (population counts 123 and diet data) used in this study were collected during the sea lion breeding season from May to 124 August because: 1) most of the data available was from this season, 2) population data from this 125 season captures the maximum number of sea lions present that year at a rookery by including 126 newborn pups, and 3) we wanted to avoid the potential confounding effects that might be 127 introduced by seasonal changes in diet when comparing diets across years and areas. 128 2.1.1 Diet data 129 Diet data refers to data on the occurrence of identified prey species from sea lion scat samples 130 collected at rookeries. Available data was used to assess diet quality, which we characterized using 131 measures of diet diversity and diet energy density. Both diet characteristics were calculated using 132 data originally reported as frequency of occurrence of prey species (FO) (27), or as an index of 133 importance (IIMP) of prey species (Gulf of California data only) (16,28). 134 To test the relationship between diet quality and population change, we ideally needed matching 135 diet data for the years with available population data. However, diet data were sparse, unevenly 136 distributed over time, and only available for certain rookeries and years. Additionally, in some 137 cases data were reported as a single mean spanning several years (e.g., Santa Barbara Island, 1981– 138 1995; Table S1). To address these limitations and make use of all available data, various data 139 processing methods were employed prior to conducting analyses. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 7 140 2.1.2 Processing diet data 141 Only prey species with FO values ≥5% were included in this dataset because: 1) this was the cut- 142 off available from most of the data, and 2) it served to highlight the main prey items. We also 143 applied a ≥5% cut-off to IIMP values to maintain consistency across the data (this is the same 144 IIMP cut-off previously used by Porras-Peters 2008). Note: while some literature uses a cut-off of 145 IIMP values ≥10% (16), we were able to apply a ≥5% IIMP value cut-off because we had access 146 to the raw diet data (F.J García-Rodriguez unpubl. data). ‘Non–identified’ species reported in the 147 IIMP data were deemed not useful for this analysis and were therefore excluded. 148 To be able to compare diet data between rookeries and years, remaining FO and IIMP values were 149 standardized to sum to 1.0 (or 100%) within each year of data by dividing each reported value by 150 the total FO or IIMP for that rookery and year. This yielded modified frequency of occurrence 151 (MFO; (29) and modified importance index (MIIMP) values which were used from this point 152 onwards to calculate diet diversity and energy density. 153 2.1.3 Calculating diet diversity 154 Two measures of diet diversity were used: 1) the total number of species recorded in the diet, and 155 2) diet diversity calculated using the Shannon Index (30) from either MFO or MIIMP values for 156 individual species, where a higher resulting H-index indicates greater species diversity. Values of 157 average diet diversity using the Shannon Index were calculated for each rookery and year when 158 data was available, which were then used in subsequent analyses (Table S2). 159 2.1.4 Calculating (weighted) diet energy density .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 8 160 The energy density of each prey species was recorded in kilojoules per gram of wet weight 161 (kJ/gww) from data obtained in Gleiber et al., 2022 (31). If an energy density value was not 162 available at the species level, an average energy density value from the species’ family (or in a few 163 cases a closely related family) was used instead to approximate energy density. Then we calculated 164 the weighted average diet energy density (to account for relative appearance of each prey species 165 in the diet) by multiplying the MFO or MIIMP (expressions of proportion in the diet) by the 166 respective energy density of that prey species (in kJ/gww). Summing these values gave an average 167 weighted diet energy density value for each rookery and year, which was used in subsequent 168 analyses (Table S2). 169 2.1.5 Population data 170 All sea lion counts from 1980–2018 from the Gulf of California rookeries were obtained from 171 (17). Counts were made from boat surveys and included numbers for each age and sex class. 172 Population counts for the Channel Islands were available from 1980–2019 and were sourced from 173 (3) for 1980–2014, from (32) for 2015, and from (6) for 2016–2019 (Table S1). These counts had 174 already been corrected for pups that were obscured from vision and for adult females that were 175 foraging during the census. At times, multiple sources reported population counts for the Channel 176 Islands, so the source with the higher counts was used if this was because a technique with 177 presumed greater accuracy was used (i.e., aerial photography counts vs. boat counts). 178 2.1.6 Calculating population change 179 To test the relationship between population and diet quality, we calculated rates of population 180 change for years with available diet data at each rookery. Using regression analysis, we estimated .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 9 181 population changes immediately associated with single or grouped years with diet data (i.e., year- 182 rookery grouping) by incorporating population counts from a set number of years before and after 183 the diet data. Since the range of years incorporated into each rookery-year diet data point varied, 184 a set of rules were established to define the number of years before and after the diet data span that 185 were included in the calculation of population change, depending on how many years of diet data 186 were included for that grouping. Further details on calculating population change provided in the 187 Supplementary Methods. In general, population trends were estimated from data that spanned the 188 period of diet data by at least an additional year on either side. 189 2.2. Grouping data 190 2.2.1 Creating Zones and Zone-era groupings 191 To investigate the relationship between population changes and diet quality, we had to ensure the 192 independence of the data points on both a temporal (both sequential and non-sequential data) and 193 geographic scale (closely related rookeries). This involved grouping rookery diet data (energy 194 density and diet diversity averages) and population data (population change averages over a set of 195 years) together into non-continuous sets of years into rookery-year groupings (more details in 196 Supplementary Methods). Then we created sets of matching population and diet data sets averaged 197 across related geographic areas (Zones) and non-continuous time periods (eras): Zone-era 198 groupings. 199 As previously noted, neither population or diet data was continuous across years, and the timing 200 of the data was not consistent across rookeries, resulting in the previously described rookery-year 201 groupings. Values for the eventual ‘Zone-era’ data groupings were created by averaging the .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 10 202 population change of each rookery-year grouping, and the respective diet quality values (energy 203 density and diversity) within a geographic Zone (described below) for each set of number of years 204 or ‘era’ with available data. 205 This grouping of data from individual rookeries into composite geographic Zones was done for 206 two main reasons: 1) to prevent potential over-representation of individual rookeries relatively 207 close to each other that could be considered common ecological units, and 2) to best deal with the 208 lack of available continuous population and diet data over time by grouping available data, thus 209 allowing us to make comparisons between different geographic Zones for similar time 210 periods/eras. Rookeries were grouped into a common Zone if they occurred in a similar geographic 211 area (<100 km away from each other) and had a similar population trajectory over time, which 212 were determined by fitting a linear regression to the total population counts of each rookery for all 213 years with available data (Fig. 2). The average annual population change (percent) was calculated 214 as the slope divided by the intercept (the predicted first year population) of the regression equation. 215 Population trajectories were therefore classified as increasing, decreasing, or inconclusive 216 according to the slope of the linear regression. 217 Previous studies have partitioned the 13 Gulf of California rookeries into three sub-populations 218 based on factors such as environmental conditions, genetic structure, and diet (17,33–36). 219 However, in our study we deliberately excluded factors related to diet, and thus created Zones 220 based only on similarities in population trajectory and geographic proximity. 221 Following this methodology, 10 Zones were created: the Channel Island rookeries were grouped 222 together into Zone 1 (Fig. 1). The available data resulted in two Zone-era groupings: Zone 1 1981– 223 1995 and Zone 1 2000–2011. The Gulf of California rookeries were grouped into 9 zones (Zones .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 11 224 2–10; Fig. 1). Data for the Gulf of California using the available FO data resulted in two era 225 groupings, 1990–2000 and 2015–2019. This resulted in 9 Zone-era groupings for the GoC using 226 FO data. The IIMP data yielded three era groupings: 1995–1996, 2002, and 2015–2019 (Table S3). 227 This resulted in 18 Zone-era groupings for GoC. Hence, this process yielded 11 Zone-era sets of 228 matched population and FO-based diet quality data for the CI and GoC combined, and an 229 additional 18 sets of data for the Gulf of California from IIMP-based diet quality. The relationship 230 between diet quality and population change was then tested on these ‘Zone-era’ groupings. 231 Fig. 1. Map of the California sea lion rookeries and designated Zones for this study. Study 232 sites included the four rookeries in the Channel Islands (a–d: San Miguel, San Nicolas, Santa 233 Barbara, San Clemente) designated as Zone 1, and the 13 rookeries along the Gulf of California 234 (e–q), and their respective Zones (2–10). Circles indicate rookeries within the indicated Zone. 235 Rookery f (San Jorge) did not have diet data available and was omitted from further analysis. For 236 reference, the distance between the northernmost rookery of Roca Consagradas (Zone 2), and the 237 southernmost rookery of Los Islotes (Zone 10) is 823 km. 238 2.3 Effects of environmental change 239 Changes in diet quality before and after an event characterized by increased sea surface 240 temperatures were explored by comparing the change in average diet energy density and diet 241 diversity in the Gulf of California before and after 2014. Sufficient post-environmental shift diet 242 data were not available for the Channel Islands post-environmental shift, so only pre- 243 environmental shift data was used to compare the diet quality between the Gulf of California and 244 the Channel Islands. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 12 245 Prey species from the diet data were grouped into 9 species categories to further describe changes 246 in diet before and after the environmental shift, and to express the ecological distribution of species 247 consumed (number of species per category). These categories were assigned based on broad 248 ecological characteristics similar to previous studies as per Trites et al., 2007 (37), and included: 249 benthic species (n = 60 prey species), crustaceans (n = 1), gadids (n = 5), lanternfish (n = 7), 250 octopus (n = 2), rockfish (n = 5), schooling fishes (n = 21), squids (n = 15), and miscellaneous (n 251 = 17) (Table S4). 252 2.4 Statistical analysis 253 To test relationships between population changes and diet quality, linear regression models were 254 fit to the data in R-Studio (version 2022.02.3). Simple linear models were used to test for 255 relationships between population change and measures of diet diversity and energy density at the 256 level of ecological Zones using Zone-era data. Additionally, since Zones consisted of varying 257 population sizes, those regression analyses incorporated median Zone population size as a 258 weighting factor. The resulting p-values and adjusted R-squared values from all linear models are 259 reported. All energy density and diet diversity values derived from MFO and MIIMP data were 260 tested for outliers using Grubb’s and Dixon’s outlier tests in R-Studio using the package “outliers”, 261 at the Zone-era grouping level. Preliminary analyses revealed no statistically significant outliers 262 in the data. 263 Two-sample t-tests assuming unequal variances were conducted using MFO and MIIMP data to 264 compare diet diversity and energy density within the Gulf of California before and after the 2014 265 environmental shift, and when comparing pre-environmental shift diets between the Gulf of 266 California and the Channel Islands. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 13 267 3. Results 268 269 3.1 Population trajectories 270 All four Channel Island rookeries (Zone 1) averaged population increases of 2-6% per year from 271 1980–2020, except for a population decrease associated with increased sea surface temperatures 272 in 2014 (Fig. 2). In contrast, the overall population in the Gulf of California decreased over the 273 same time-period, although trends differed at individual rookeries. While most rookery 274 populations showed a decline, San Esteban and Los Islotes rookeries (Zones 6 and 10) showed 275 increasing population trajectories (although the San Esteban population increase was not 276 statistically significant). As opposed to the Channel Islands, sea lion populations in the Gulf of 277 California did not seem to be collectively affected by the 2014 warming event (Fig. 2). 278 Fig. 2. California sea lion population trends within Zones (1980–2020). Data shows total sea 279 lion counts for each Zone with data from 1980–2020. Rookery names within each Zone are labeled 280 in the bottom right of each panel. Zones composed of multiple rookeries show the sum of the 281 population totals in those rookeries. Solid lines represent statistically significant regressions and 282 dotted lines represent regressions that were not statistically significant. Red data points in Zone 1 283 represent population declines after the 2014 warming event (2015–2019); these data were not 284 included in the overall regression. 285 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 14 286 3.2 Diet quality and population changes 287 Diet diversity (using the Shannon Index) did not correlate significantly with rate of population 288 change when examined at the level of Zone-eras (p = 0.430 using MFO data and 0.622, using IIMP 289 data; S3 Fig., left panels). Nor were there significant relationships between diet energy densities 290 and rates of population change (p = 0.804 using MFO data and p = 0.128 using IIMP data; S3 Fig., 291 right panels). 292 3.3 Diet quality in the Channel Islands vs. the Gulf of California 293 At the Channel Islands, California sea lions primarily consumed 23 main prey species, whereas in 294 the Gulf of California, they consumed 88 main species (using MFO >5% data; Figs. 3 and 4). Jack 295 mackerel, Chub mackerel, Pacific hake, and Californian anchovy coincided as top species 296 consumed by sea lions in both the Channel Islands and the Gulf of California. 297 Fig. 3. Prevalence of California sea lion prey species in the Channel Islands. Bars represent 298 the total number of occurrences (out of 41 possible occurrences, equivalent to years) of each prey 299 species from frequency of occurrence data from 1980–2011; that is, the total number of years 300 where each prey species was present in the diet. All 23 species with FO ≥5% are listed. The 301 species’ common name is listed when available, although some were identified only at the family 302 or genus level. 303 Fig. 4. Prevalence of California sea lion prey species in the Gulf of California. Bars represent 304 total prey species occurrences in the diet per year (total number of years where each prey species 305 was present in the diet), from available diet data from 1980–2019 for all Gulf of California .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 15 306 rookeries (Zones 2–10) from either frequency of occurrence (blue bars) or index of importance 307 (orange bars). The common name of all 88 species (or species groups) from both MFO and MIIMP 308 data is listed when available (some species were identified only at the family or genus level). 309 Sea lions in the Channel Islands ate mainly schooling fishes and squid (36% and 21% of diet 310 respectively; data from 1981–2011), whereas sea lions in the Gulf of California mainly ate multiple 311 benthic species (41 species; 47% of diet) and schooling fishes (19 species; 22% of diet; data from 312 1990–2019; Fig. 5, left panels), with schooling fishes having a higher average energy density than 313 benthic species (Fig. 6). 314 Fig. 5. Diet composition by prey species categories before and after 2014 from frequency of 315 occurrence data. Pie chart slices represent the proportion by each species category. White 316 numbers represent the number of species in the diet from each category. Diet composition data 317 from the Channel Islands is from 1980–2011 (no comparable diet data available after 2011). Diet 318 composition data for the Gulf of California before the 2014 environmental shift is from 1990–2000 319 and from 2015–2019 after 2014. 320 Fig. 6. Average energy density of each prey species category consumed by California sea 321 lions. Colours correspond to species categories illustrated in Fig. 6. Bars represent average energy 322 densities (kJ/gww; mean value ± standard error) from all species present in the diet data from each 323 category ordered from highest (lanternfish) to lowest (octopus) value. 324 Mean diet energy density for sea lions in the Channel Islands (5.43kJ/gww) was comparable to the 325 Gulf of California (5.32 kJ/gww) (Table 1 and Fig. 7). Mean diet diversity using the Shannon 326 Index was lower in the Channel Islands than in the Gulf of California (1.83 vs. 2.04; Table 1 and .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 16 327 Fig. 7) and had higher variability among the Gulf of California Zones (range: 0.79–3.26) than 328 among the Channel Islands (range: 1.34–2.35). .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 17 329 Table 1. Diet quality and population trajectory of each of the Gulf of California Zones and the Channel Islands from frequency 330 of occurrence data. The mean and standard deviation of the range of the annual weighted diet energy density (kJ/gww) and of the diet 331 diversity (Shannon Index) for sea lion diets is shown (incorporating individual prey species). ‘n’ represents the number of years with 332 data points used to calculate the corresponding mean and range values. Population trajectories in bold represent statistical significance. 333 Zones 2, 8 & 9 are omitted due to lack of available FO data. Energy density (kJ/gww) Diet diversity (Shannon Index)Zone: rookery Population trajectory Mean ±SD Range Mean ±SD Range n 1: San Miguel Increasing 5.43 ±0.94 4.14–7.25 1.79 ±0.94 1.34–2.35 10 1: San Nicolas Increasing 5.63 ±0.22 5.26–5.94 1.75 ±0.23 1.46–1.96 8 1: San Clemente Increasing 5.16 ±0.84 3.79–6.05 1.94 ±0.84 1.67–2.21 7 1: Santa Barbara Increasing 5.48 N/A 1.87 N/A 1 Channel Islands average Increasing 5.43 ±0.20 3.79–7.27 1.83 ±0.12 1.34–2.35 4 3: Isla Lobos Decreasing 5.66 ±0.66 5.19–6.13 1.93 ±0.66 1.81–2.04 2 4: Machos, Cantiles, Granito Decreasing 4.90 ±0.35 4.63–5.39 2.05 ±0.35 0.79–3.26 4 5: Rasito Decreasing 5.04 ±2.45 4.31–5.77 2.34 ±1.03 2.20–2.48 2 6: San Esteban Increasing 5.90 ±2.83 5.85–5.96 2.04 ±0.08 1.81–2.27 2 7: San Pedro Mártir Decreasing 5.32 ±2.50 5.27–5.37 2.14 ±0.07 2.07–2.21 2 10: Los Islotes Increasing 5.06 ±2.17 4.88–5.48 1.90 ±0.28 1.56–2.40 4 Gulf of California average Decreasing 5.32 ±0.39 4.31–6.13 2.04 ±0.15 0.79–3.26 6 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 18 335 Fig. 7. Average annual diet diversity and energy density by Zones from frequency of 336 occurrence data. Diet diversity (top panel) and energy density (bottom panel) values were based 337 on all available Zone and year groupings from MFO (1981–2019; Zones 2, 8 & 9 are omitted due 338 to lack of available FO data). Diet diversity values were calculated using the Shannon Index. Box 339 limits represent the first, mean, and third quantile values, box whiskers represent the range of 340 values. Bar widths are proportional to the number of data points in each Zone. The circle represents 341 the energy density value of 7.25 in Zone 1 (San Miguel, 2005) which was considered an outlier 342 according to Grubb’s test. 343 The lowest diet diversity (0.79) occurred within Zone 4 (Granito, Cantiles, Machos) in 1996, which 344 was heavily influenced by the Granito rookery where just one main species, largehead hairtail 345 (Trichiurus lepturus), was consumed that year. Interestingly, the highest diversity recorded among 346 all locations (3.26) occurred in 2018 also within Zone 4 where data were exclusively from Granito 347 and showed that that year, sea lions consumed a total of 31 main species. 348 The mean diet energy density in the Channel Islands was 5.43 ±0.2 kJ/gww, with a surprisingly 349 small overall variation considering that two data points had anomalously high or low energy 350 densities (7.25 kJ/gww in San Miguel, 2005 and 3.79 kJ/gww in San Clemente, 1982). The species 351 that contributed the most to the average energy density in the diet primarily belonged to the 352 schooling fishes category (Fig. 6), and included jack mackerel, Pacific mackerel, and northern 353 anchovy. San Miguel Island in particular had years when sea lion diets had above-average energy 354 densities (S1 Fig.) with a higher-than-normal contribution from two species of schooling fishes, 355 Pacific sardine (2002–2005) and herring (2005). .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 19 356 Although the overall diet energy density for sea lions in the Gulf of California was not statistically 357 different than for those in the Channel Islands, there was significant variability between years and 358 within Zones in the Gulf of California, demonstrating differences in the species that contributed 359 the most to the average diet energy densities. The highest diet energy density in the Gulf of 360 California in any year occurred in Zone 3 (Isla Lobos, 1995, 6.13 kJ/gww), mainly due to the high 361 energy density of Pacific anchoveta (a schooling fish) and largehead hairtail (a miscellaneous fish) 362 (Table 1 and S2 Fig.). However, diets in Zone 6 (San Esteban) had the highest mean energy density 363 overall (5.90 kJ/gww; 1995 and 1996). For both Zones 6 and 7 (San Pedro Mártir), the overall 364 energy density of the diet largely reflected a high contribution from lanternfish, followed by 365 largehead hairtail, Californian anchovy, and chub mackerel (S2 Fig.). Interestingly, in 1996 366 lanternfish was largely replaced with other species in the diet in both Zones 6 and 7 with little 367 effect on the mean diet energy density. Diets of sea lions feeding in Zone 4 (Granito, Cantiles, 368 Machos) had the lowest mean energy density (4.90 kJ/gww), which included a high proportion of 369 ‘other’ species outside of the top 17 for most years (1996, 2016, 2018) (S2 Fig.). 370 3.4 Effect of environmental changes on diet quality in the Gulf of 371 California 372 373 From 2014 to 2016, the Gulf of California experienced unusually high sea surface temperatures. 374 During this period, the proportions of schooling fish and squid in the diet of California sea lions 375 decreased significantly, from 27% to 16% and from 7% to 1%, respectively, while the proportion 376 of benthic species increased from 36% to 56% (Fig. 5). These changes marked a shift from high- 377 energy schooling fish to lower-energy-density benthic species, resulting in a significant reduction .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 20 378 in the overall average energy density of their diet to 4.69 ± 0.35 kJ/gww (two-tailed p = 0.045; 379 Table 2, Fig. 8, bottom panel). 380 381 Table 2. Diet quality by geographic area before and after 2014. Total number of diet prey 382 species, mean diet diversity using the Shannon Index, and mean weighted diet energy density. Geographic area & era Number of prey species in diet Diet diversity (Shannon Index) (mean ±SD) Energy density (kJ/gww; mean ±SD) Channel Islands before 2014 23 1.83 ±0.12 5.42 ±0.36 Gulf of California before 2014 51 1.92 ±0.49 5.22 ±0.53 Gulf of California after 2014 65 2.30 ±0.57 4.69 ±0.35 383 384 Fig. 8. Average diet diversity from the Shannon Index (top panel) and average energy density 385 (bottom panel) from frequency of occurrence data before and after 2014. Box limits represent 386 averaged data (first, median and third quantiles ± standard error, ‘x’ represents mean value) from 387 all Channel Island and Gulf of California groupings before and after 2014 (no data for Channel 388 Islands after 2014). Diet data is based on rookery-year groupings. No significant differences were 389 found between mean diversity values between geographic areas before 2014, nor in the Gulf of 390 California between eras. There was a statistically significant decrease in the average energy density 391 in the Gulf of California after 2014. 392 393 Diet diversity calculated using the Shannon Index showed no significant difference in mean values 394 before (1.92 ±0.49) and after 2014 in the Gulf of California (2.30 ±0.57; two-tailed p = 0.245; 395 Table 2, Fig. 8, top panel). However, there was an increase in the number of prey species consumed .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 21 396 (from 51 to 65 species), and an overall increase in the average number of prey species consumed 397 per rookery after 2014 (S4 Fig.). Before 2014, sea lions consumed 9 prey species on average 398 within rookeries (range 5–15 species), increasing to an average of 16 species per rookery (range 399 7–26 species) after 2014 (S5 Fig.; data unavailable for Zones 8 and 9). Equally notable was that 400 ~50% of the species consumed throughout the Gulf of California after 2014 were not present in 401 the diet prior to this time (S6 Fig.) ––in other words, the sea lions did not simply add 14 more 402 species to their diet but made a fundamental shift in the species they consumed. Despite the overall 403 dietary shifts in prey quality observed after 2014, there were no apparent differences in the rates 404 of population changes between these two eras (two tailed p = 0.984). 405 406 4. Discussion 407 Previous dietary studies on California sea lions in the Gulf of California have tended to focus on 408 detailing differences in the main prey species consumed, or describing feeding behaviours at 409 various rookeries (16,18). Only one study has indirectly explored the broad relationship between 410 population changes and diet within the Gulf of California, finding no significant relationships 411 between these variables (17). Our study is the first to assess California sea lion diet quality using 412 available long-term summer data on a finer geographic scale for the Gulf of California (1990– 413 2019) compared to the Channel Islands (1980–2014), to explore its relationship to differences in 414 population trajectories. 415 Our results demonstrate substantial differences in the diversity and type of prey species consumed 416 by sea lions in these two areas, but do not show any significant relationships between measures of 417 diet quality (diet energy density or diet diversity) and long-term rates of population change. Results .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 22 418 also demonstrate that following the significant increase in sea surface temperatures that occurred 419 in 2014, sea lions in central regions of the Gulf of California consumed a greater number of prey 420 species that had an overall lower diet energy density. However, this shift did not result in an overall 421 population decline throughout the Gulf, as occurred in the Channel Islands. These findings 422 underscore the importance of considering the environmental heterogeneity at the different regions 423 throughout the Gulf of California, which can heavily influence California sea lion population 424 dynamics at local levels. 425 426 4.1 The role of diet diversity 427 An ideal diet for California sea lions is one that allows them to meet their nutritional needs to grow 428 and reproduce by feeding on sufficiently available prey species. However, our results illustrate 429 that the exact nature of such a diet appears to vary depending on the characteristics of the 430 ecosystem, making single indicators such as diet diversity difficult to interpret. Animals may 431 choose to forage on fewer, energy-rich prey species (a low diversity, high energy density diet), 432 which may reflect either a high degree of prey selectivity or a less biodiverse ecosystem. In more 433 diverse ecosystems, foraging on a greater combination of species of different sizes and nutritional 434 profiles might prove to be an optimal strategy and may even be required to fulfill nutritional 435 requirements (38). Different marine mammal species have exhibited switches to lower quality diets 436 (lower energy density prey) during environmental challenges that have been characterized by 437 either decreases (22,39) or increases (40,41,26) in diet diversity. 438 Using the total number of prey species as a measure of diet diversity in our study revealed a striking 439 difference in diets between regions. The sea lions at the Channel Islands consistently consumed 440 23 primary prey species during the summer (1981–2011, Fig. 3) while those in the Gulf of .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 23 441 California consumed 88 primary prey species (1990–2019; Fig. 4) that varied between Zones (S2 442 Fig.). There was a lower number of prey species observed in the Channel Islands, despite the fact 443 that diet diversity in this region using the Shannon Index is highest during summer (20). These 444 regional differences in the number of prey items may mean that the ideal prey species were not 445 consistently available for sea lions in the Gulf, or it may alternatively illustrate that there were 446 more prey options available to facilitate diet adaptability in some Zones. 447 448 Overall, the Shannon Index and the total number of species consumed illustrate that the diet has 449 been relatively consistent over time within the Channel Islands (Fig. 7, Table 1). In comparison, 450 there was greater variability in diet composition in the Gulf of California both between Zones and 451 over time within each Zone (S2 Fig.). For example, the number of species consumed per rookery 452 ranged from 5–26 species between Zones (S5 Fig.), indicating greater resource heterogeneity 453 and/or apparent dietary flexibility. 454 455 4.2 The role of diet energy density 456 Regardless of its relationship to diet diversity, the energy density of a diet is an important 457 characteristic to consider when assessing diet quality. As each prey species differs in macronutrient 458 composition and therefore in energy density (kJ/gww), diets based upon prey species that are more 459 energy-rich can be considered “higher quality” as they are more likely to meet the nutritional 460 requirements of individuals, allowing populations to grow. Measuring diet energy density can 461 provide important insight into the nutritional status of a population and the drivers of population 462 change. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 24 463 On average, the diet energy density across the Gulf of California did not differ significantly from 464 that of the Channel Islands, such that energy density did not explain broad differences in population 465 trajectories. Contrary to expectations, we found that regions and periods when the diet had the 466 highest energy density were not necessarily associated with years of greatest population growth 467 within Zones. Alternate analyses (i.e., using finer data at the rookery-year scale and using changes 468 in pup numbers as a more immediate indicator of changes in population demographics) also did 469 not reveal any relationships between diet energy density and population changes. Even the diets 470 with the highest energy densities were associated with both increasing and decreasing population 471 trajectories, indicative of the lack of an overall simple relationship between diets and populations 472 over time. 473 The highest mean diet energy density of all Zones (including the Channel Islands) occurred in the 474 Gulf of California Zone 6 (San Esteban rookery), the largest rookery in terms of population size 475 in the Gulf, where the sea lion population only showed a potentially increasing trend, while Zone 476 3 (Isla Lobos) had the second highest diet energy density but a significantly decreasing population. 477 Furthermore, the only significantly increasing population in the Gulf of California (Zone 10, Los 478 Islotes rookery) had a mean diet energy density that was comparable to the median diet energy 479 density for all Zones (Table 1). While there is evidence from other pinniped studies to support the 480 hypothesized link between diet energy density and population growth (24,40), it is possible that 481 differences in diet quality in our study regions were not great enough to be the primary population 482 drivers across the regions, except perhaps in cases with extremely low-quality diets (e.g., the 483 decreasing population in Zone 4). .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 25 484 An important factor to consider when exploring the relationship between populations and diet 485 quality is that, as demonstrated by our results, the Channel Islands and the Gulf of California are 486 fundamentally different oceanographic systems with different population and diet dynamics. In 487 the Channel Islands, most of the diet energy density consistently comes from schooling fishes (S1 488 Fig.), whereas there is considerable variability in the diet of sea lions in the Gulf of California, 489 which is made up of different combinations of benthic species (41 different benthic species in 490 total) making up their main source of energy from food (S2 Fig.), even though this prey category 491 has a lower average energy density than schooling fishes. In fact, the greater predictability of the 492 prey available to sea lions in the Channel Islands than those in the Gulf of California may be a 493 more important contributor to population dynamics, rather than differences in prey energy density 494 per se. 495 The heterogeneity in diet between rookeries throughout the Gulf of California raises questions 496 about the specific trade-offs and foraging strategies among sea lions breeding at different 497 rookeries. In the Galápagos Islands, individual foraging strategies of Galápagos sea lions 498 (Zalophus wollebaeki) influence the coping abilities of their population with evidence that some 499 foraging strategies may be more advantageous than others during environmental changes (42). In 500 contrast to pelagic foragers, benthic foraging Galápagos sea lions appear to be less affected by 501 increased water temperatures, despite consuming prey that are lower in energy density. Such 502 environmentally dependent fitness trade-offs could also be at play in the Gulf of California sea 503 lion populations, which should be explored through further research on individual foraging 504 strategies. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 26 505 4.3 The effects of environmental changes on diet quality 506 From 2013–2015, a large-scale phenomenon of increased sea surface temperatures known as “The 507 Blob” was first documented in Alaska and then traveled south along the Eastern Pacific. In 2015– 508 2016, The Blob coincided with a strong El Niño event, further intensifying the effect of increased 509 water temperatures (43). In the Channel Islands, El Niño events are known to cause population 510 declines (with rapid subsequent recovery), and are associated with a decrease in consumption of 511 energy-rich schooling fish (13). The Blob’s effects on the Channel Island populations were first 512 seen in 2013 through increased pup mortality (20). 513 Unlike the Pacific coast, sea lions within the Gulf of California are not affected in the same way 514 by warming events in the Pacific Ocean, such as El Niño. Instead, they appear to be affected by 515 local oceanographic processes within the Gulf’s sub-regions (5). Overall, although the increased 516 water temperatures in the Gulf after 2014 were characterized by an increase in the mean and range 517 of diet diversity within the Gulf of California (expressed using the Shannon Index), this change 518 was not significant (Fig. 8, top panel). However, there was an increase in the total number of 519 species present in the diet (from 51 to 65 species, Table 2). In addition to the increase in the number 520 of species consumed, the proportion of each diet species category changed (Fig. 5), as well as the 521 identity of the species within those categories. For example, around 50% of the species consumed 522 after 2014 were not present in the diet in previous years. These species consisted mostly of new 523 benthic and lanternfish species (Figs. 7 and S6). 524 This increase in diet diversity was accompanied by a significant decrease in the overall average 525 diet energy density throughout the Gulf of California (Fig. 8, bottom panel). This trend was driven 526 by central rookeries (Zones 4 and 5) where there was a decrease in the proportion of energy-rich .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 27 527 schooling fishes after 2014 (Fig. 9). Although there was also an increase after 2014 in the overall 528 number of high-energy lanternfish species in the diet (from 1 to 6 species; Fig. 5), there was no 529 significant change in their proportion (6% vs. 5%). This could suggest sea lions attempted to 530 continue to meet their total energy requirements by increasing the diversity of energy-rich 531 lanternfish species in their diet. 532 Fig. 9. Average energy density of California sea lion diets at rookeries in the Gulf of 533 California before and after 2014 from frequency of occurrence data. Only rookeries within 534 Zones with matched data before and after 2014 are included in this Fig., their Zone number is 535 shown in brackets. Gulf of California Zones are ordered geographically from North (Z4) to South 536 (Z10). Zones that are not shown here lacked data after 2014 and were excluded from this Fig.. 537 538 However, this pattern of decreased diet energy density after 2014 (mainly due to a decrease in 539 energy-rich schooling fishes and lanternfish) was not consistent across all rookeries. A decrease 540 in diet energy density after 2014 occurred at Rasito (Zone 5), which had a lack of energy-rich 541 lanternfish and Jack mackerel in 2016 compared to 1996 (S2 Fig.). Conversely, within Zone 10 542 (Los Islotes rookery), an increase in lanternfish in 2019 was largely responsible for the overall 543 increase in diet energy density after 2014 (Fig. 9). 544 The observed differences in the changes in diet energy density post-2014 among different Zones 545 could be due to a difference in the availability of prey species in the various regions in the Gulf of 546 California. In some Zones this could lead to California sea lions having to adopt atypical, lower- 547 energy diets. For example, the unusual dominance of ‘other’ species in the diet in Los Islotes in 548 2015 may reflect a loss of ideal primary prey due to increased water temperatures (S2 Fig.), .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 28 549 resulting in them consuming a higher number of prey species (S5 Fig.) to maintain the same level 550 of dietary energy density from prey (~5kJ/gww). 551 Previous studies have demonstrated how acute environmental changes and subsequent prey 552 availability shifts can affect marine mammal population growth. For example, ringed seals (Phoca 553 hispida) in western Hudson Bay switched to a more diverse diet that had a lower energy density 554 due to decreases in the availability of their main prey (sand lance) triggered by changes in the 555 seasonal breaking up of sea ice (40). As a result, body condition of individual seals was greatly 556 reduced, and population declines ensued. 557 In the Channel Islands, models predict that every 1C increase in surface temperature could 558 decrease the population growth rate of California sea lions in the U.S. by 7% (44). Of note, during 559 years of increased sea surface temperatures (2013–2015), (20) reported how sea lion diet 560 composition in the Channel Islands decreased in epipelagic species (schooling fish), and increased 561 in benthic and demersal species. A similar phenomenon appears to be occurring in the northern 562 and central regions of the Gulf of California where California sea lion pup birth rates declined as 563 sea surface temperature anomalies exceeded 1C (17). However, population growth in the southern 564 regions of the Gulf was not affected by increased sea surface temperatures (45), although pup 565 abundance and body condition at Los Islotes did decrease during 2014 and 2015, and adult females 566 were away from the rookery for longer periods than normal (46). Lactating females appeared to 567 have had to forage further away from the rookery, which cost them more time and energy. Thus, 568 warmer sea surface temperatures affected both the diet and foraging behaviour of female California 569 sea lions in the south (Los Islotes, Zone 10), which may have affected pups during the lactation 570 period. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 29 571 4.4 Environmental heterogeneity and its implications for species 572 management 573 The variable diet quality and population trends detailed in our study suggest that sea lions at 574 different rookeries, even those just within the Gulf of California, cannot be viewed nor managed 575 as a homogeneous group. The Gulf of California is known to have considerable environmental 576 heterogeneity (47–50), which may influence both the quality of sea lion diets as seen in our study 577 and, ultimately, predator-prey dynamics at the rookery scale. It has been suggested, for example, 578 that compared to the northern and central regions of the Gulf, the greater diversity of prey species 579 present in the south may buffer rookeries like Los Islotes against detrimental environmental 580 changes (50,51). Having access to a greater diversity of prey species would allow sea lions to 581 compensate for prey that may no longer be available. 582 Prey availability and abundance in the Gulf of California varies by region and is not as consistent 583 or as predictable as in the California Current System. Such variability may mask the ability to 584 identify simple relationships between diet and population growth, such as those shown in other 585 Eastern Pacific ecosystems like the Channel Islands or Alaska (23,32,37). This variability may 586 also underly differences noted by others in terms of genetic differences between California sea 587 lions, their foraging areas, and the oceanographic conditions they experience (4,17,34). 588 Understanding diet and population dynamics in the Gulf of California may therefore require a 589 more detailed understanding of sea lion prey dynamics, foraging strategies, and localized 590 oceanographic changes. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 30 591 The Mexican government deems the California sea lion sub-populations in need of special 592 protection, and recognizes the need to recover and conserve the populations in Mexico’s rookeries 593 (52). Our study highlights the variation in the diets, population trajectories, rookery sizes, and 594 oceanographic dynamics within the Gulf of California, suggesting that each rookery population 595 faces different sets of challenges that impact their reproduction and survival rates in different ways. 596 However, current management and surveillance programs (especially in the central region) do not 597 seem sufficient to monitor and assess how sea lion numbers are affected. More rigorous monitoring 598 is needed not only to understand changes in prey species, but also other short-term factors affecting 599 sea lion numbers such as entanglements in fishing gear, shootings, and contaminants to name a 600 few (53). Such anthropogenic factors may also contribute to the lack of a direct relationship 601 between diet and population trends in the northern and central Gulf of California (17). Overall, 602 understanding the complex dynamics affecting each sub-population in both the short-term and 603 long-term is essential to effectively manage the protection and conservation of California sea lions 604 in the Gulf of California. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 31 605 Acknowledgements 606 607 The present work was part of Ana Lucía Pozas Franco’s Master’s thesis. Funding consisted of an 608 NSERC Discovery Grant to Dr. David Rosen. Thank you to Dr. Miram Gleiber for sharing key 609 energy density data unpublished at the time, Dr. Tony Orr for initial input and guidance, Eric Lee 610 for help categorizing prey species, and Pedro González Espinosa for help in creating the Zone 611 map. 612 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 32 613 References 614 615 1. Lowry M, Maravilla-Chávez O. Recent abundance of California sea lions in western Baja 616 California, Mexico and the United States. 6th Calif Isl Symp. 2005;485–97. 617 2. Peterson RS, Bartholomew GA. The Natural History and Behavior of the California Sea Lion. 618 American Society of Mammalogists; 1967. (Special Publication No. 1). 619 3. Lowry M, Melin S, Laake J. Breeding season distribution and population growth of California 620 sea lions, Zalophus californianus, in the United States during 1964-2014 (NOAA-TM-NMFS- 621 SWFSC-574). U.S. DEPARTMENT OF COMMERCE; 2017. (NOAA Technical 622 Memorandum NMFS). 623 4. Adame K, Elorriaga-Verplancken FR, Beier E, Acevedo-Whitehouse K, Pardo MA. The 624 demographic decline of a sea lion population followed multi-decadal sea surface warming. Sci 625 Rep. 2020;10(1):1–15. 626 5. Pelayo-González L, González-Rodríguez E, Ramos-Rodríguez A, Hernández-Camacho CJ. 627 California sea lion population decline at the southern limit of its distribution during warm 628 regimes in the Pacific Ocean. Reg Stud Mar Sci. 2021;48:102040. 629 6. Lowry M, Jaime EM, Moore JE. Abundance and Distribution of Pinnipeds at the Channel 630 Islands In Southern California, Central and Northern California, and Southern Oregon During 631 Summer 2016-2019 (NOAA-TM-NMFS-SWFSC-656) [Internet]. National Oceanic and 632 Atmospheric Administration; 2021. (NOAA Technical Memorandum NMFS). Available 633 from: https://doi.org/10.25923/6qhf-0z55 634 7. Hernández-Camacho CJ, Pelayo-González L, Rosas-Hernández MP. California Sea Lion 635 (Zalophus californianus, Lesson 1828). In: Ecology and Conservation of Pinnipeds in Latin 636 America. Springer Nature Switzerland AG 2021; 2021. p. 119–43. 637 8. Loughlin TR, York AE. An accounting of the sources of Steller sea lion, Eumetopias jubatus, 638 mortality. Mar Fish Rev. 2000;40–5. 639 9. Read AJ, Drinker P, Northridge S. Bycatch of marine mammals in U.S. and global fisheries. 640 Conserv Biol. 2006;20(1):163–9. 641 10. Reeves R, McClellan K, Werner T. Marine mammal bycatch in gillnet and other entangling 642 net fisheries, 1990 to 2011. Endanger Species Res. 2013 Mar 21;20(1):71–97. 643 11. Stroud R, Roffe T. Causes of death in marine mammals stranded along the Oregon coast. J 644 Wildl Dis. 1979;15(1):91–7. 645 12. Lowry M, Stewart B, Heath CB, Yochem P, Francis JM. Seasonal and annual variability in 646 the diet of California sea lions Zalophus californianus at San Nicolas island, California, 1981 647 -86. Fish Bull. 1991;2(89):331–6. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 33 648 13. Melin SR, Orr AJ, Harris JD, Laake JL, Delong RL, Gulland FMD, et al. Unprecedented 649 mortality of California sea lion pups associated with anomalous oceanographic conditions 650 along the central California coast in 2009. Calif Coop Ocean Fish Investig Rep. 2010;51:182– 651 94. 652 14. Aurioles-Gamboa D, Fox C, Sinsel F, Tanos G. Prey of the California sea lion (Zalophus 653 californianus) in the Bay of La Paz, Baja California Sur, Mexico. J Mammal. 1984;65(3):519– 654 21. 655 15. Aurioles-Gamboa D, Zavala-González A. Ecological factors that determine distribution and 656 abundance of the California sea lion (Zalophus californianus) in the Gulf of California. Cienc 657 Mar. 1994;20:535–53. 658 16. García-Rodríguez FJ, Aurioles-Gamboa D. Spatial and temporal variation in the diet of the 659 California sea lion (Zalophus californianus) in the Gulf of California, Mexico. Fish Bull. 660 2004;102(1):47–62. 661 17. Pelayo-González L, Aurioles-Gamboa D, Godínez-Reyes C, Rubio-Rodríguez U, Hernández- 662 Camacho CJ, Gallo-Reynoso JP, et al. Effect of environmental variables on the number of 663 births at California sea lion (Zalophus californianus) rookeries throughout the Gulf of 664 California, Mexico. Aquat Conserv Mar Freshw Ecosyst. 2021;(March 2020):1–19. 665 18. Porras-Peters H, Aurioles-Gamboa D, Cruz-Escalona VH, Koch PL. Trophic level and overlap 666 of sea lions (Zalophus californianus) in the Gulf of California, Mexico. Mar Mammal Sci. 667 2008;24(3):554–76. 668 19. Melin SR, DeLong RL, Siniff DB. The effects of El Niño on the foraging behavior of lactating 669 California sea lions (Zalophus californianus californianus) during the nonbreeding season. 670 Can J Zool. 2008;86(3):192–206. 671 20. Lowry M, Nehasil S, Moore J. Spatio-temporal diet variability of the California sea lion 672 (Zalophus californianus) in the southern California Current Ecosystem. Mar Ecol Prog Ser. 673 2022;692:1–21. 674 21. Fritz L, Brost B, Laman E, Luxa K, Sweeney K, Thomason J, et al. A re-examination of the 675 relationship between Steller sea lion (Eumetopias jubatus) diet and population trend using data 676 from the Aleutian Islands. Can J Zool. 2019;97(12):1137–55. 677 22. Merrick RL, Chumbley MK, Byrd GV. Diet diversity of Steller sea lions (Eumetopias jubatus) 678 and their population decline in Alaska: a potential relationship. Can J Fish Aquat Sci. 1997; 679 23. Rosen DAS, Trites A. Pollock and the decline of Steller sea lions: Testing the junk-food 680 hypothesis. Can J Zool. 2000;78(7):1243–50. 681 24. Trites A, Donnelly CP. The decline of Steller sea lions Eumetopias jubatus in Alaska: A review 682 of the nutritional stress hypothesis. Mammal Rev. 2003;33(1):3–28. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 34 683 25. Trites A. Behavioral Insights into the Decline and Natural History of Steller Sea Lions. In: 684 Ethology and Behavioral Ecology of Otariids and the Odobenid. Springer; 2021. p. 489–519. 685 26. Jory C, Lesage V, Leclerc A, Giard J, Iverson S, Bérubé M, et al. Individual and population 686 dietary specialization decline in fin whales during a period of ecosystem shift. Sci Rep. 687 2021;11(1):1–14. 688 27. Lance MM, Orr AJ, Riemer SD, Weise MJ, Laake JL. Pinniped Food Habits and Prey 689 Identification Techniques Protocol (AFSC Processed Report 2001-04) [Internet]. Alaska 690 Fisheries Science Center; 2001 p. 1–35. Available from: https://apps- 691 afsc.fisheries.noaa.gov/Publications/ProcRpt/PR%202001-04.pdf 692 28. Garcia-Rodriguez FJ, De La Cruz-Agüero J. A comparison of indexes for prey importance 693 inferred from otoliths and cephalopod beaks recovered from pinniped scats. J Fish Aquat Sci. 694 2011;6(2):186. 695 29. Bigg M, Perez M. A review of methods of diet analysis in piscivorous marine mammals. In: 696 Marine Mammals and Fisheries [Internet]. George Allen and Unwin, London; 1985. Available 697 from: 698 https://www.researchgate.net/publication/237043776_A_review_of_methods_of_diet_analys 699 is_in_piscivorous_marine_mammals 700 30. Shannon C, Weaver W. The Mathematical Theory of Communication. Univ Ill Press. 701 1949;(Urbana). 702 31. Gleiber MR, Hardy NA, Roote Z, Morganson CJ, Krug-Macleod A, George I, et al. Pelagic 703 Species Trait Database [Internet]. Scholars Portal Dataverse, V1. 2022 [cited 2022 Sep 13]. 704 Available from: https://doi.org/10.5683/SP3/0YFJED 705 32. Lowry M, Nehasil SE, Jaime EM. Distribution of California sea lions, Northern elephant seals, 706 Pacific harbor seals, and Steller sea lions at the Channel Islands during July 2011-2015 707 [Internet]. National Oceanic and Atmospheric Administration; 2017 p. 64. (NOAA Technical 708 Memorandum NMFS). Report No.: NOAA-TM-NMFS-SWFSC-578. Available from: 709 doi:10.7289/V5/TM-SWFSC-578 710 33. González-Suárez M, Flatz R, Aurioles-Gamboa D, Hedrick PW, Gerber LR. Isolation by 711 distance among California sea lion populations in Mexico: Redefining management stocks. 712 Mol Ecol. 2009;18(6):1088–99. 713 34. Schramm Y, Mesnick SL, De La Rosa J, Palacios DM, Lowry M, Aurioles-Gamboa D, et al. 714 Phylogeography of California and Galápagos sea lions and population structure within the 715 California sea lion. Mar Biol. 2009;156(7):1375–87. 716 35. Szteren D, Aurioles-Gamboa D. Ecological regionalization of Zalophus californianus 717 rookeries, as a tool for conservation in the Gulf of California. Cienc Mar. 2011;37(3):349–68. 718 36. Ward EJ, Chirakkal H, González-Suárez M, Aurioles-Gamboa D, Holmes EE, Gerber L. 719 Inferring spatial structure from time-series data: Using multivariate state-space models to .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 35 720 detect metapopulation structure of California sea lions in the gulf of California, Mexico. J Appl 721 Ecol. 2010;47(1):47–56. 722 37. Trites A, Calkins DG, Winship A. Diets of Steller sea lions (Eumetopias jubatus) in Southeast 723 Alaska, 1993-1999. Fish Bull. 2007;105(2):234–48. 724 38. Raubenheimer D, Simpson SJ. Integrating nutrition: a geometrical approach. In Springer; 725 1999. p. 67–82. 726 39. Choy ES, Giraldo C, Rosenberg B, Roth JD, Ehrman AD, Majewski A, et al. Variation in the 727 diet of beluga whales in response to changes in prey availability: Insights on changes in the 728 Beaufort Sea ecosystem. Mar Ecol Prog Ser. 2020;647:195–210. 729 40. Chambellant M, Stirling I, Ferguson SH. Temporal variation in western Hudson Bay ringed 730 seal Phoca hispida diet in relation to environment. Mar Ecol Prog Ser. 2013;481:269–87. 731 41. Páez-Rosas D, Moreno-Sánchez X, Tripp-Valdez A, Elorriaga-Verplancken FR, Carranco- 732 Narváez S. Changes in the Galapagos sea lion diet as a response to El Niño-Southern 733 Oscillation. Reg Stud Mar Sci. 2020;40:101485. 734 42. Schwarz JFL, DeRango EJ, Zenth F, Kalberer S, Hoffman JI, Mews S, et al. A stable foraging 735 polymorphism buffers Galápagos sea lions against environmental change. Curr Biol. 736 2022;32(7):1623-1628.e3. 737 43. Tseng YH, Ding R, Huang XM. The warm Blob in the northeast Pacific - The bridge leading 738 to the 2015/16 El Niño. Environ Res Lett. 2017;12(5):054019. 739 44. Laake JL, Lowry M, DeLong RL, Melin SR, Carretta JV. Population growth and status of 740 California sea lions. J Wildl Manag. 2018;82(3):583–95. 741 45. Molina-Chávez VG. Caracterización de los movimientos en el mar de hembras de lobo marino 742 de California (Zalophus californianus) [MSc Thesis]. [La Paz]: Centro Interdisciplinario de 743 Ciencias Marinas; 2022. 744 46. Cruz-Vallejo RA. Abundancia y amplitud alimentaria del lobo marino de California de Los 745 Islotes, Bahía de La Paz, México y su relación con el ambiente (2013-2018) [MSc Thesis]. 746 Instituto Politécnico Nacional, Centro Interdisciplinario de Ciencias Marinas; 2020. 747 47. Thomson DA, Gilligan MR. Rocky-shore fishes. In: In Island Biogeography in the Sea of 748 Cortés II. Oxford University Press.; 1983. 749 48. Lavin, Marinone. An Overview of the Physical Oceanography of the Gulf of California. 750 Nonlinear Process Geophys Fluid Dyn. 2003;173–204. 751 49. Brusca RC, Findley LT, Hastings PA, Hendrickx ME, Cosio JT, van der Heiden AM. 752 Macrofaunal diversity in the Gulf of California. Biodivers Ecosyst Conserv North Mex. 753 2005;179. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 36 754 50. Álvarez-Borrego S. Physical, chemical and biological oceanography of the Gulf of California. 755 In: The Gulf of California: Biodiversity and Conservation [Internet]. The University of 756 Arizona Press; 2010. p. 24–48. Available from: 757 https://www.academia.edu/33338546/Marine_Mammals_of_the_Gulf_of_California_An_Ov 758 erview_of_Diversity_and_Conservation_Status 759 51. Durán-Campos E, Coria-monter E, Monreal-gómez MA, Salas-de-león DA, Mazatlán UA, 760 Ciencias ID, et al. Impact of " the Blob " 2014 and 2019 in the sea surface temperature and 761 chlorophyll- a levels of the Gulf of California : a satellite-based study. Lat Am J Aquat Res. 762 2022;50(3):479–91. 763 52. NOM-059-SEMARNAT. Norma Oficial Mexicana NOM-059-SEMARNAT-2010, 764 Protección Ambiental-Especies Nativas de México de Flora y Fauna Silvestres-Categorías De 765 Riesgo y Especificaciones Para Su Inclusión, Exclusión o Cambio-Lista de Especies en Riesgo 766 Prefacio. 2010. 767 53. Hernández-Camacho CJ, González-López I, Pelayo-González L, Aurioles-Gamboa D, López- 768 Greene E, Rosas-Hernández MP. Effective management of the national park Espíritu Santo, 769 through the governance, planning, and design of an integral strategy for Los Islotes. Socio- 770 Ecol Stud Nat Prot Areas. 2020;(1):679–704. 771 772 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 37 773 Supplementary Methods 774 775 776 Calculating population change rules 777 In instances where a rookery had diet data for 3 or more continuous years, population data from 778 one year before to one year after the diet data years were incorporated into the population change 779 calculation. If the number of continuous years with diet data was less than three, then the 780 population change was calculated from two years before and after the interval (or single year) of 781 diet data. This yielded a single rate of population change over the matching diet data interval. In 782 cases where population changes using the previously mentioned rules were calculated to be 783 greater than ±20% (an unrealistic growth rate under normal breeding conditions), four years on 784 either side of the diet data were incorporated into the calculation to obtain a more realistic rate of 785 population change. 786 Estimating missing population totals 787 Rookeries and years with available diet data were paired with available population totals. In 788 some cases, population totals had to be estimated for years that lacked data. For the Channel 789 Island rookeries, pup counts were available for years with missing population totals, and were 790 therefore used to estimate totals by extrapolating from the linear relationship between pup counts 791 and total population counts across all years. For the Gulf of California rookeries, missing 792 population numbers were estimated by extrapolating from a linear regression performed on all 793 available population data for individual rookeries because additional years with pup count data 794 were not available. The regression equation was then used to estimate population numbers for 795 years lacking counts, and the rate of population change over the period of interest was calculated .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 38 796 from this mixed data set. Inaccurate extrapolations were avoided by only using regressions that 797 spanned intervals with actual counts. 798 In cases where rookeries had one year of diet data and where the associated population data 799 range (when incorporating the ±2 years rule for population change) overlapped with the 800 population data range for another year of diet data, and one of those diet data years did not have 801 +2 years of data after (due to it being the latest population year with data), those years of diet 802 data were grouped. For example, diet data from the Granito rookery from 2016 and 2018 were 803 combined into one grouping, and the years 2014–2018 were used to calculate population change 804 since 2018 was the latest year with population data. 805 806 Calculating rookery-year groupings 807 808 We matched the available diet data at the rookery level with a rate of population change value 809 calculated to correspond to the specific year or group of consecutive years with available diet data. 810 Ideally, continuous diet data for all years and rookeries would have been available, and matching 811 data groupings would have been strategically chosen. However, diet data were patchy in terms of 812 both years and rookeries. As a result, sequential data points were compiled from each rookery to 813 form specific ‘year-rookery groupings’ and were treated as independent data points used in all 814 subsequent analyses (Table S2). 815 Most rookeries in the Channel Islands had diet data available over several consecutive years, which 816 were grouped according to the continuity of the data (e.g., San Miguel 2009–2011). Diet data from 817 several rookeries in the Channel Islands were already averaged over several years (Table S2). In 818 these instances, those year-rookery groupings were kept and used when calculating corresponding .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 39 819 population changes. Within the Gulf of California, most years with diet data were single isolated 820 years that varied by rookery (e.g., Los Islotes 1990, 2000, 2015 and 2019; Rasito 1996 and 2016, 821 etc.). In cases where two consecutive years of diet data were available, we averaged the diet data 822 and grouped them to form one rookery-year grouping (e.g., San Esteban 1995–1996). In cases 823 where non-continuous years with diet data were close in time such that their population change 824 calculations overlapped, we also combined them to form a single rookery-year grouping (e.g., Los 825 Islotes 2015, 2019). 826 Figure S1. Energy density of diets of California sea lions in the Channel Islands (Zone 1). 827 Average energy density and energetic content contributions (average weighted energy density) of 828 the top 17 prey species to the total energetic content of the diet for rookeries and years with 829 available frequency of occurrence data in the Channel Islands. ‘Other’ category represents all other 830 species in the diet beyond the top 17. 831 832 Figure S2. Energy density of diets of California sea lions in the Gulf of California (Zones 3– 833 5). Average energy density and energetic content contributions (average weighted energy density) 834 of the top 17 prey species to the total energetic content of the diet for rookeries and years with 835 available frequency of occurrence data in the Gulf of California. ‘Other’ category represents all 836 other species in the diet beyond the top 17. 837 838 Figure S3. Population changes and diet quality for California sea lions from frequency of 839 occurrence and index of importance data. The data represents values from Zone-era groupings. 840 Diet diversity values were calculated using the Shannon Index from frequency of occurrence data 841 (top-left panel) and index of importance data (bottom-left panel). Panels on the right include diet .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 40 842 energy density values calculated using frequency of occurrence (top) and index of importance data 843 (bottom). Regression analysis of data weighted by rookery population size indicated no statistical 844 relationships. 845 846 Figure S4. Average number of prey species per California sea lion rookery before and after 847 2014. Bars represent averaged data (mean value ± standard error). Average number of species 848 is based on frequency of occurrence data. There was no data after 2014 available for the Channel 849 Islands. 850 851 Figure S5. Average number of prey species per California sea lion rookery before and after 852 2014 from frequency of occurrence data. Blue bars represent each of the Channel Islands before 853 2014, each rookery with available data in the Gulf of California is shown before 2014 (green bars) 854 and after 2014 (orange bars). Zone number is shown in brackets. 855 856 Figure S6. Total number of prey species consumed by California sea lions in the Gulf of 857 California before and after 2014 from frequency of occurrence diet. The bar on the left 858 represents prey species before 2014, the bar on the right represents prey species after 2014. Green 859 bars show number of species present in both eras, yellow bar shows species only present before 860 2014, and orange bar shows number of species present only after 2014. 861 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 41 862 Table S1. Data sources of meta-analysis of California sea lion population and diet. Details 863 include the rookery, data year and season, diet index or total population estimate (unless stated), 864 and Table in source publication where relevant data is found. (FO; frequency of occurrence data, 865 IIMP; index of importance data). Source Rookeries Year(s) of data Season Data type/diet index Table Channel Islands San Miguel 1971–1991 (pup counts) 1992–2014 July or August Population (total live) Table 2 San Nicolas 1991–2008 2009, 2010 (pup counts) 2011–2014 July Population (total live) Table 2 San Clemente 1981–2014 July or August Population (total live) Table 2 Lowry et al. (2017a) Santa Barbara 1983–1985 (pup counts) 1986–2008 2009, 2010 (pup counts) 2011–2014 July Population (total live) Table 2 Lowry et al. (1991) San Nicolas 1981–1986 (grouped) June and August FO Table 1, 2 Lowry and Carretta (1999) San Nicolas, Santa Barbara, San Clemente 1981–1995 Multiple months, summer (Santa Barbara) FO Table 3 A. Curtis unpubl. data San Nicolas, San Clemente 1981–1986 Summer FO N/A Lowry et al. (2017b) Channel Islands 2015 July Population (total counts) Table 3 Lowry et al. (2021) Channel Islands 2016–2019 July or August Population (total live) Table 1 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 42 866 867 Source Rookeries Year(s) of data Season Data type/diet index Table Orr et al. (2011) San Miguel 2002–2006 March– July FO Table 3 S. R. Melin et al. (2012) San Miguel 2000–2003, 2005, 2009– 2011 June–Sept FO Table 5 S. R. Melin et al. (2010) San Miguel 2000, 2001, 2002, 2004, 2005 and 2009 July–early August FO Table 1 Gulf of California Pelayo-González, González- Rodríguez, et al. (2021) 1–13 (not all rookeries have data for all years) 1980–2019 June or July Population (total) Raw data García-Rodríguez (1995) 13 1990 February– September FO Table 4 Source Rookeries Year(s) of data Season Data type/diet index Table Gulf of California F.J Garcia- Rodriguez unpubl. data 5 1995 September FO & IIMP N/A 3, 4, 6, 9, 10 1995 June 3, 4, 5, 8, 9, 10 1996 May F.J Garcia- Rodriguez unpubl. data, Cardenas-Palomo (2003) 13 2000 May 2000– April 2001 FO & IIMP Cuadro 1, Anexo 3 Porras-Peters (2004) and Porras- Peters et al. (2008) 1, 3, 5, 7–13 2002 Summer IIMP Figure 4 (Appendix II) 13 2015 July 4 2016 October Pelayo-González et al., (2021) 5 2016 Unknown FO Raw data .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 43 868 Gulf of California rookeries: 1: Rocas Consagradas, 2: San Jorge (no data), 3: Isla Lobos, 4: 869 Granito, 5: Cantiles, 6: Machos, 7: El Partido, 8: El Rasito, 9: San Esteban, 10: San Pedro Mártir, 870 11: San Pedro Nolasco, 12: Farallón de San Ignacio, 13: Los Islotes. 871 Table S2. Raw paired diet and population data for California sea lions by rookery-year 872 grouping. Rookery-year groupings Zone Index Average diet diversity Weighted energy density (kJ/gww) Population change Average energy density (kJ/gww) Pup change (%) San Miguel 2000–2006 1 FO 1.70 5.76 5.03% 5.6 3.2% San Miguel 2009–2011 1 FO 2.00 4.68 7.80% 5.1 11.6% San Nicolas 1981–1986 (grouped) 1 FO 1.92 5.66 -2.92% 5.36 -4.8% San Nicolas 1981–1986 1 FO 1.68 5.65 -2.92% 5.36 -4.8% San Nicolas 1981–1995 1 FO 1.95 5.46 9.47% 5.85 32.0% Santa Barbara 1981–1995 (grouped) 1 FO 1.87 5.48 9.95% 5.85 24.8% San Clemente 1981–1995 (grouped) 1 FO 1.81 5.59 4.51% 5.74 7.7% San Clemente 1981–1986 1 FO 1.96 5.09 -1.61% 5.50 -1.9% Los Islotes 1990 10 FO 2.40 4.59 2.43% 4.7 5.12% San Pedro Mártir 1995–96 7 FO 2.14 5.3 -1.53% 4.93 -0.18% 6 2016 October IIMP 8 2016 October FO & IIMP 4 2018 August 4 2018 July 5 2018 July 13 2019 August FO .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 44 Rookery-year groupings Zone Index Average diet diversity Weighted energy density (kJ/gww) Population change Average energy density (kJ/gww) Pup change (%) San Esteban 1995–96 6 FO 2.04 5.9 7.29% 5.08 -1.52% Rasito 1996 5 FO 2.48 5.77 -7.60% 5.15 2.31% Machos 1995 4 FO 1.96 5.74 -1.17% 5.29 4.37% Cantiles 1995– 96 4 FO 1.36 4.51 6.15% 5.05 6.76% Granito 1995– 96 4 FO 0.90 5.34 3.70% 5.15 12.74% Isla Lobos 1995–96 3 FO 1.93 5.13 -2.03% 5.57 5.16% Los Islotes 2000 10 FO 2.03 4.69 7.37% 5.00 3.26% Rasito 2016 5 FO 2.20 4.31 7.25% 4.31 13.52% Machos 2016 4 FO 2.01 4.66 -11.65% 4.92 17.32% Cantiles 2016 4 FO 2.70 4.42 7.09% 4.58 7.92% Granito 2016, 2018 4 FO 3.02 4.85 7.09% 5.15 12.71% Los Islotes 2015, 2019 10 FO 1.59 5.19 -1.18% 5.1 8.44% San Pedro Mártir 1995–96 7 IIMP 1.526 5.86 -1.5% 5.58 -0.2% San Esteban 1995–96 6 IIMP 1.315 6.29 7.3% 6.48 -1.5% Rasito 1995–96 5 IIMP 1.682 5.19 -11.2% 5.50 3.9% Machos 1995 4 IIMP 1.528 5.94 -1.2% 5.75 4.4% Cantiles 1995– 96 4 IIMP 0.931 4.27 6.1% 4.85 6.8% Granito 1995– 96 4 IIMP 0.561 5.30 3.7% 5.37 12.7% Isla Lobos 1995–96 3 IIMP 1.479 5.19 -2.0% 4.47 5.2% Los Islotes 2002 10 IIMP 1.848 4.7 4.4% 5.08 -0.2% .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 45 Rookery-year groupings Zone Index Average diet diversity Weighted energy density (kJ/gww) Population change Average energy density (kJ/gww) Pup change (%) San Esteban 2002 6 IIMP 1.503 5.3 9.2% 5.42 16.3% Rasito 2002 5 IIMP 0.062 5.43 5.5% 4.85 12.9% San Pedro Mártir 2002 7 IIMP 0.685 3.10 -0.7% 4.98 -6.4% Cantiles 2002 4 IIMP 1.096 3.81 -3.6% 3.81 -8.8% Isla Lobos 2002 3 IIMP 1.326 4.71 4.5% 5.01 -2.1% San Pedro Nolasco 2002 8 IIMP 1.874 4.99 -1.4% 5.11 -2.1% Partido 2002 5 IIMP 1.454 6.81 -7.9% 5.29 -3.9% Rocas Consagradas 2002 2 IIMP 1.060 5.33 5.4% 4.63 5.4% Farallón de San Ignacio 2002 9 IIMP 2.386 6.20 -3.8% 4.94 -5.2% Los Islotes 2015 10 IIMP 0.957 4.38 1.3% 5.03 8.2% Rasito 2016 5 IIMP 1.921 4.17 7.2% 4.31 13.5% Granito 2016, 2018 4 IIMP 1.300 4.59 6.6% 4.84 12.7% Cantiles 2016, 2018 4 IIMP 1.433 5.15 7.1% 4.90 7.9% Los Islotes 2019 10 IIMP 1.684 6.10 -3.7% 5.09 8.6% 873 874 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 46 875 Table S3. Raw paired diet and population data for California sea lions by Zone-era grouping. Zone-era groupings Era Index Average diet diversity Population change Weighted average energy density (kJ/gww) Median population Zone 1 – Channel Islands 1981– 1995 FO 1.87 2.7% 5.49 9,216 Zone 1 – Channel Islands 2000– 2011 FO 1.85 6.4% 5.22 44,720 Zone 3 – Isla Lobos 1995– 1996 FO 1.93 -2.0% 5.13 2,822 Zone 4 – Machos, Cantiles, Granito 1995– 1996 FO 1.41 2.9% 5.20 1,355 Zone 4 – Machos, Cantiles, Granito 2016, 2018 FO 2.58 0.8% 4.64 696 Zone 5 – Rasito 1996 FO 2.48 -7.6% 5.77 362 Zone 5 – Rasito 2016 FO 2.20 7.2% 4.31 308 Zone 6 – San Esteban 1995– 1996 FO 2.04 7.3% 5.90 7,171 Zone 7 – San Pedro Mártir 1995– 1996 FO 2.14 -1.5% 5.32 1,963 Zone 10 – Los Islotes 1990 FO 2.21 4.9% 4.64 347 Zone 10 – Los Islotes 2015, 2019 FO 1.59 -1.2% 5.19 538 Zone 2 – Rocas Consagradas 2002 IIMP 1.06 5% 5.33 839 Zone 3 – Isla Lobos 1995– 1996 IIMP 1.48 -2% 5.19 2,822 Zone 3 – Isla Lobos 2002 IIMP 1.33 4% 4.71 1,897 Zone 4 – Machos, Cantiles, Granito 1995– 1996 IIMP 1.01 3% 5.17 1,355 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 47 Zone-era groupings Era Index Average diet diversity Population change Weighted average energy density (kJ/gww) Median population Zone 4 – Machos, Cantiles, Granito 2002 IIMP 1.10 -4% 3.81 1,090 Zone 4 – Machos, Cantiles, Granito 2016, 2018 IIMP 1.20 7% 4.87 729 Zone 5 – Rasito, Partido 1995– 1996 IIMP 1.68 -11% 5.19 366 Zone 5 – Rasito, Partido 2002 IIMP 0.76 -1% 6.12 507 Zone 5 – Rasito, Partido 2016 IIMP 1.30 7% 4.17 308 Zone 6 – San Esteban 1995– 1996 IIMP 1.32 7% 4.99 7,171 Zone 6 – San Esteban 2002 IIMP 1.45 9% 6.81 6,334 Zone 7 – San Pedro Mártir 1995– 1996 IIMP 1.53 -2% 5.86 7,171 Zone 7 – San Pedro Mártir 2002 IIMP 0.69 -1% 3.10 2,405 Zone 8 – San Pedro Nolasco 2002 IIMP 1.87 -1% 4.99 937 Zone 9 – Farallón de San Ignacio 2002 IIMP 2.39 -4% 6.20 643 Zone 10 – Los Islotes 2002 IIMP 1.85 4% 4.72 404 Zone 10 – Los Islotes 2015 IIMP 1.92 1% 4.38 538 Zone 10 – Los Islotes 2019 IIMP 1.68 -4% 6.10 659 876 877 878 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 48 879 Table S4. List of 114 diet prey species (scientific and common name) consumed by California 880 sea lions in the Gulf of California (MEX) or the Channel Islands (USA) showing their average 881 energy density, category assigned, and country where data was collected. Some species had 882 the same scientific name, but different common names depending on the region. Categories: B: 883 benthic species, C: crustaceans, G: gadids, L: lanternfish, O: octopus, R: rockfish, SF: schooling 884 fish, S: squid, M: miscellaneous. Scientific name Common name Energy Density (kJ/gww) Category Country Abraliopsis affinis Squid 4.40 S MEX Abraliopsis species Squids 4.40 S USA Anisotremus davidsonii Xantic sargo 4.88 B MEX Apogon retrosella Barspot cardinalfish 4.70 B MEX Argentina sialis North-Pacific argentine 3.57 M MEX Atherinops species Topsmelt silverside 6.20 SF MEX Atherinopsis californiensis Jack silverside 6.20 M MEX Aulopus Royal flagfin 4.43 B MEX Aulopus bajacali Eastern Pacific flagfin 4.43 B MEX Balistes polylepis Finescale triggerfish 3.84 B MEX Bodianus diplotaenia Mexican hogfish 3.84 B MEX Brosmophycis marginata Red brotula 3.39 B MEX Calamus brachysomus Pacific porgy 7.45 B MEX Caulolatilus princeps Ocean whitefish 7.45 B MEX Ceratoscopelus townsendi Dogtooth lampfish 7.16 L MEX Cetengraulis mysticetus Pacific anchoveta 6.01 SF MEX Chromis punctipinnis Blacksmith damselfish 4.68 B USA Citharichthys species Flatfish 3.33 B MEX Clupea pallasii Pacific herring 7.51 SF USA Coelorinchus scaphopsis Shoulderspot grenadier 5.10 G MEX Cololabis saira Pacific saury 7.50 M USA Cynoscion reticulatus Shorefish 7.99 B MEX Decapodiformes Superorder of squids 4.60 S USA Diaphus theta California headlightfish 9.88 L MEX Diplectrum macroposoma Mexican sand perch 4.50 B MEX Diplectrum pacificum Inshore sand perch 4.03 B MEX Diplectrum species Sandperch 5.02 B MEX Doryteuthis opalescens Opalescent inshore squid 3.70 S USA Dosidicus gigas Humboldt squid 5.39 S MEX Engraulidae Anchovies 6.17 SF MEX .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 49 Scientific name Common name Energy Density (kJ/gww) Category Country Engraulis mordax Californian anchovy 6.70 SF MEX Engraulis mordax Northern anchovy 6.80 SF USA Girella nigricans Rudderfish 4.40 M MEX Gonatopsis borealis Boreopacific armhook squid 4.20 S USA Gonatus berryi Berry armhook squid 5.02 S MEX Gonatus onyx Clawed armhook squid 5.86 S USA Gonatus species Armhook squid 5.90 S USA Haemulidae species Grunt fish 4.88 B MEX Haemulon californiensis Yellowspotted grunt 4.88 M MEX Haemulon flaviguttatum Greybar grunt 4.88 M MEX Haemulon sexfasciatum Scaled-fin grunt 4.88 SF MEX Haemulon species Californian salema 4.88 SF MEX Haemulopsis leuciscus Raucous grunt 4.88 B MEX Haemulopsis species Grunt fish 4.88 B MEX Hemanthias peruanus Splittail bass 4.50 SF MEX Hemanthias species Sea bass 4.50 B MEX Hermosilla azurea Zebra perch 4.40 SF MEX Holacanthus passer King angelfish 7.45 B MEX Icelinus tenuis Spotfin sculpin 5.82 B MEX Lepophidium prorates Prowspine cusk eel 3.39 B MEX Lestidiops species Barracudina 4.30 SF MEX Leuroglossus stilbius California smoothtongue 3.90 M USA Loliolopsis diomedeae Dart squid 3.75 S MEX Lycodes cortezianus Bigfin eelpout 7.60 B USA Merluccius productus North Pacific hake 4.20 G MEX Merluccius productus Pacific Hake 4.20 G USA Merluccius species Hake 4.07 G MEX Micropogonias ectenes Slender croaker 7.99 B MEX Micropogonias species Croaker 7.99 B MEX Myctophidae Lanternfish 7.62 L MEX Nannobrachium species Lanternfish 7.63 L MEX Octopus rubescens East Pacific red octopus 3.30 O USA Octopus species Octopus 3.40 O USA, MEX Oegopsida Pelagic squid 4.50 S MEX Onychoteuthidae Hooked squid family 5.40 S USA .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 50 Scientific name Common name Energy Density (kJ/gww) Category Country Onychoteuthis borealijaponicus Boreal clubhook squid 5.48 S USA Ophidion scrippsae Basketweave cusk-eel 3.39 B MEX Ophidion species Cusk-eels 3.39 B MEX Ophistonema species Herrings 7.47 SF MEX Orthopristis reddingi Bronze-striped grunt 4.88 SF MEX Oxylebius pictus Painted greenling 4.44 B MEX Paralabrax clathratus Kelp bass 4.45 M MEX Paralabrax species Rock bass 4.45 B MEX Paralichthys californicus California halibut 3.81 B MEX Physiculus nematopus Charcoal mora 4.00 G MEX Physiculus species Codling 4.00 G MEX Pleuroncodes planipes Pelagic red crab (lobster) 6.70 C USA Pontinus furcirhinus Red scorpionfish 3.19 B MEX Pontinus species Scorpionfish 3.19 B MEX Porichthys notatus Plainfin midshipman 3.36 B MEX Porichthys species Midshipman 3.36 B MEX Prionotus species Searobin 4.63 B MEX Prionotus stephanophrys Lumptail searobin 4.63 B MEX Pronotogrammus eos Bigeye bass 4.45 B MEX Pronotogrammus multifasciatus Threadfin bass 4.45 B MEX Sarda lineolata Pacific bonito 7.04 SF MEX Sardinops caeruleus California pilchard 7.47 SF MEX Sardinops sagax South American pilchard 7.50 SF MEX Sardinops sagax Pacific sardine 7.50 SF USA Scomber japonicus Chub mackerel 6.80 SF MEX Scomber japonicus Pacific mackerel 6.80 SF USA Scopelengys tristis Pacific blackchin 7.62 M MEX Scorpaenidae Scorpionfish 3.19 M MEX Sebastes exsul Buccaneer rockfish 5.51 R MEX Sebastes jordani/species Rockfish 5.60 R USA Sebastes macdonaldi Mexican rockfish 5.51 R MEX Sebastes species Rockfish 5.60 R MEX Selar crumenophthalamus Bigeye scad 6.27 SF MEX Serranus aquidens/aequidens Deepwater serrano 4.45 B MEX .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint 51 Scientific name Common name Energy Density (kJ/gww) Category Country Specieshyraena argentea Pacific barracuda 3.36 M MEX Stenobrachius leucopsarus Northern lampfish 9.70 L USA Strongylura exilis Californian needlefish 6.20 SF MEX Symbolophorus californiensis Bigfin lanternfish 7.07 L MEX Symphurus fasciolaris Banded tongue fish 4.00 B MEX Symphurus species Tongue fish 4.00 B MEX Synodus lucioceps California lizardfish 4.43 B MEX Synodus species Lizardfish 4.25 B MEX Trachurus species Jack mackerel 6.30 SF MEX Trachurus symmetricus Pacific jack mackerel 6.27 SF MEX Trachurus symmetricus Jack mackerel 6.30 SF USA Trichiurus lepturus Largehead hairtail 4.76 M MEX Trichiurus nitens Pacific cutlassfish 5.05 M MEX Triphoturus mexicanus Mexican lampfish 7.07 L MEX Zaniolepis species Combfish 7.60 B USA 885 886 887 888 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 26, 2025. ; https://doi.org/10.1101/2025.04.23.650199doi: bioRxiv preprint