Current status of the brown trout (Salmo trutta) populations within eastern Pyrenees genetic refuges-SALMONID SYMPOSIUM GIRONA-

Since the end of the 20th century, some headwaters of rivers in the eastern Pyrenees have been designated as genetic refuges to protect remaining native brown trout (Salmo trutta) diversity. The declaration was based on limited or no evidence of genetic impact from released non-native Atlantic hatchery fish. Hatchery releases were completely banned into the genetic refuges, but pre-existing fishing activities were maintained. Specific locations in each refuge have been monitored every 2–3 trout generations to update genetic information to accurately assess the contribution of these reservoirs to the preservation of native brown trout gene pools. This work updates genetic information to year 2014 in three of these locations (in Ter, Freser and Flamisell rivers). Previous studies identified hatchery introgressed populations within refuges and suggested discrepancies between the underlying intention of the genetic refuges and the gene pools detected. Therefore, we also examined genetic divergences among locations inside refuge river segments. Combined information at five microsatellite and the lactate dehydrogenase C (LDH-C*) loci showed reduced but significant temporal native allele frequency fluctuations in some of the above specific locations that did not modify overall levels of local diversity and river divergences. Bayesian clustering analyses confirmed the presence of differentiated native units within each genetic refuge. Some locations of the Freser River within the genetic refuge area showed high hatchery impact of non-native fish (over 20%). We discuss additional local actions (releases of native fish, selective removals and fishery reinforcement with sterile individuals) to improve the conservation objective of genetic refuges


25
Preservation of genetic diversity within species maintains their evolutionary potential 26 and thereby the long-term conservation of the species (Ryman et al. 1995;Hurt & 27 Hedrick 2004;Utter 2004). However, rates of anthropogenic hybridization and 28 introgression are increasing dramatically worldwide because of intentional 29 translocations of organisms and habitat modifications by humans (Allendorf et al. 2001; 30 Champagnon et al. 2012;Chunco 2014). This situation is especially significant in game 31 species, where captive-bred animals derived from native, alien, or hybrid stocks are 32 often released in large numbers into the wild with the intention of reinforcing exploited 33 populations (Mamuris et al. 2001;Negro et al. 2001;Vernesi et al. 2003;Barilani et al. (Lewin et al. 2006;Naish et al. 2007) due to exotic 41 species introductions, overfishing and/or release of non-native stocks (Cowx & findings indicated that the establishment of a genetic refuge did not reduce the average 79 regional abundance of the foreign stock alleles, though the policy of genetic refuges 80 controlled the increase of introgression from 1993 to 1999, and maintained major 81 trends in the pattern of population structure. Similar results were reported from wild 82 French trout populations in the Mediterranean Alpine rivers where genetic refuges 83 policies were also implemented (Caudron et al. 2011(Caudron et al. , 2012.

84
At local scale, some monitored reference locations increased abundance of hatchery 85 alleles as a consequence of upstream migration of admixed or released hatchery fish 86 from non-refuge areas (Araguas et al. 2008). Therefore, the monitoring of genetic   each locus in each population was estimated using GENEPOP 4.0 (Rousset, 2008   176 were conducted for each tested K value (from "1" to "number of sampling sites within 177 genetic refuge + 1"). The most likely K value was estimated following Evanno et al.

209
Estimates of hatchery introgression (q) calculated by STRUCTURE software from 210 genotypes at the 5 microsatellite loci had higher background noise than the ones 211 computed from 9 loci (Table 2). For instance, the pure native brown trout population in exceeded the average q-value (Fig. 2), clearly confirming the existence of fish with 218 some degree of hatchery ancestry. In addition, the LDH-C genotypes confirmed 219 hatchery impact in all these locations. Thus, the combined information from the 5 220 microsatellite and the LDH-C* locus permitted to identify pure and hybrid trout    where native fish predominated in all locations (Table 2). In the Freser River, results 237 indicated moderate (10-30%) impact of hatchery releases throughout the basin, but a 238 population of introduced hatchery fish was present in the tributary Segadell stream at also had higher diversity levels, due to the significant genetic divergence among 244 hatchery and native Iberian gene pools. In spite of that, there were not significant 245 temporal changes on overall diversity indexes and gene diversity despite the LDH-C*90 246 allele being not detected in 1993 collections (Table 3).  (Table S2) and 251 AMOVA analysis (Table 4). When BA location was excluded in the AMOVA analysis, the percentage of genetic differentiation was reduced from 24.41% to 14.34 % (the 253 same value was obtained when PAR14 was excluded from the analysis). This  (Table 4). Over all sampled 262 collections of the year 2014, AMOVA analyses indicated that the lowest but highly 263 significant percentage of variation was among locations within genetic refuges (Table   264 4).

265
FCA analysis basically grouped individuals by basin (Fig. 3). However, individuals from placed close to the BA fish. STRUCTURE plots confirmed a higher introgression 268 impact in the Freser refuge than in Ter and Flamisell refuges (Fig. 2). Several fish 269 collected in the Freser River basin showed large proportion of ancestry of the BA 270 cluster. Freser and Ter refuges, which are hydrographically closer (Fig. 1), were also 271 genetically more similar in comparison with Flamisell refuge.

356
The establishment of genetic refuges in Mediterranean rivers has not resulted in a 357 rapid and significant decrease of hatchery alleles at regional scale (Araguas et al.

700
Codes are defined in Table 1

715
Flamisell refuges and Baga hatchery. Location codes are shown on Table 1.