Madagascar is a marine biodiversity hotspot. A recent checklist recorded 1689 marine or transitional water fish species, 2.5% being endemic. To date, studies in this country were mostly focused on adult fishes using morphological-based identification. The early life stages of fishes remain largely understudied. The present work aimed to improve knowledge of fish biodiversity in Madagascar by focusing on post-larval reef fishes and settled juveniles in seagrass meadows of southwest Madagascar by using either species identification keys or DNA barcoding. Up to 119,500 individuals were collected, and 1096 individuals were successfully barcoded. We identified 387 species—85 through their morphology (with 58 unsuccessfully sequenced) and 302 by using CO1 barcoding corresponding to 302 barcode index numbers (BINs). This study added 27 new BINs for the BOLD database, 120 new for Madagascar, but only 159 were assigned a precise species name. By referring to the updated checklist of Madagascar fishes, 10 new species were detected for Madagascar. This number will probably increase when all the barcoded specimens become assigned to precise species names. These preliminary findings stress our poor knowledge of marine fish biodiversity in Madagascar and demonstrate the relevance of DNA barcoding in improving this knowledge.

Post-larval prediction is important, as post-larval supply allows us to understand juvenile fish populations. No previous studies have predicted post-larval fish species richness and abundance combining molecular tools, machine learning, and past-days remotely sensed oceanic conditions (RSOCs) obtained in the days just prior to sampling at different scales. Previous studies aimed at modeling species richness and abundance of marine fishes have mainly used environmental variables recorded locally during sampling and have merely focused on juvenile and adult fishes due to the difficulty of obtaining accurate species richness estimates for post-larvae. The present work predicted post-larval species richness (identified using DNA barcoding) and abundance at 2 coastal sites in SW Madagascar using random forest (RF) models. RFs were fitted using combinations of local variables and RSOCs at a small-scale (8 d prior to fish sampling in a 50 × 120 km² area), meso-scale (16 d prior; 100 × 200 km²), and large-scale (24 d prior; 200 × 300 km²). RF models combining local and small-scale RSOC variables predicted species richness and abundance best, with accuracy around 70 and 60%, respectively. We observed a small variation of RF model performance in predicting species richness and abundance among all sites, highlighting the consistency of the predictive RF model. Moreover, partial dependence plots showed that high species richness and abundance were predicted for sea surface temperatures <27.0°C and chlorophyll a concentrations <0.22 mg m^-3. With respect to temporal changes, these thresholds were solely observed from November to December. Our results suggest that, in SW Madagascar, species richness and abundance of post-larval fish may only be predicted prior to the ecological impacts of tropical storms on larval settlement success.

Madagascar harbors a rich marine biodiversity, yet detailed knowledge of its fish species remains limited. Of the 1689 species listed in 2018, only 22% had accessible cytochrome oxidase I (COI) sequences in public databases. In response to growing pressure on fishery resources, this study aims to strengthen biodiversity monitoring tools. Its objectives were to enrich the COI database for Malagasy marine fishes, create the first 12S reference library, and evaluate the taxonomic resolution of different 12S metabarcodes for eDNA analysis, namely MiFish, Teleo1, AcMDB, Ac12S, and 12SF1/R1. An integrated approach combining morphological, molecular, and phylogenetic analyses was applied for specimen identification of fish captured using various types of fishing gear in Toliara and Ranobe Bays from 2018 to 2023. The Malagasy COI database now includes 2146 sequences grouped into 502 Barcode Index Numbers (BINs) from 82 families, with 14 BINs newly added to BOLD (The Barcode of Life Data Systems), and 133 cryptic species. The 12S library comprises 524 sequences representing 446 species from 78 families. Together, the genetic datasets cover 514 species from 84 families, with the most diverse being Labridae, Apogonidae, Gobiidae, Pomacentridae, and Carangidae. However, the two markers show variable taxonomic resolution: 67 species belonging to 35 families were represented solely in the COI dataset, while 10 species from nine families were identified exclusively in the 12S dataset. For 319 species with complete 12S gene sequences associated with COI BINs (Barcode Index Numbers), 12S primer sets were used to evaluate the taxonomic resolution of five 12S metabarcodes. The MiFish marker proved to be the most effective, with an optimal similarity threshold of 98.5%. This study represents a major step forward in documenting and monitoring Madagascar’s marine biodiversity and provides a valuable genetic reference for future environmental DNA (eDNA) applications.

Understanding the interannual effect of various environmental factors on biodiversity distribution is fundamental for developing biological monitoring tools. The interannual variability of environmental factors on presettlement fish assemblages (PFAs) has been so far under investigated, especially in Madagascar. Numerous explanatory variables including local hydro-dynamic conditions recorded during the sampling night, characteristics of the benthic substrate and remotely sensed oceanic conditions (RSOC) were used to explain the spatio-temporal variability of PFAs in southwestern Madagascar. Gradient forest analyses were used to hierarchically classify the effect of these explanatory variables on the PFAs for two sites and during two different recruitment seasons. RSOC variables appeared to better explain the PFAs than the local variable and the characteristics of the benthic substrate. The PFAs caught in water masses with coastal characteristics were better explained than those with open water characteristics. This spatial variability is hypothesised to be linked to differences in feeding conditions among water masses. The gradient forest analyses also highlighted the complexity of predicting PFAs as the species for which abundances were better explained by RSOC variables varied between years. This interannual variability was mainly explained by the interannual variation of chlorophyll a (Chl a) concentration, wind and surface current, with better prediction obtained during the year with high Chl a values associated with high averaged sea surface temperature. These findings suggest the importance of forecasting Chl a concentrations, taking into account the impact of tropical storms and climate variability in order to predict PFAs in the future.