Climate change and human disturbance have significantly impacted the Caspian Sea (CS) over recent decades (Karpinsky et al. 2005). In addition to occurring over a range of spatial and temporal scales, these effects alter marine ecosystem structure and function under a combination of local and global pressures (Mackinson and Daskalov 2007). Sea surface level (SSL) and rising sea water temperatures are two environmental factors driving changes in the Caspian ecosystem that affect the fish abundance and community structure (Beyraghdar Kashkooli et al. 2017, Fazli et al. 2021, 2022a).

The ichthyofauna of the whole Caspian basin includes 159 species (Naseka and Bogutskaya 2009), of which 119 species belong to 63 genera that were reported in the southern CS basin (Esmaeili et al. 2018). In this region, R. kutum, C. aurata, and other bony fishes such as Cyprinus carpio, R. rutilus, Vimba vimba and Alburnus chalcoides are the most frequently caught species using the beach seine method (Fazli et al. 2022a). During the last three decades, the annual catch of bony fishes ranged from 10910 t to 23800 t. The species R. kutum and C. aurata are among the most important commercial fish, comprising 60.5% and 21.7% of the total catch, respectively, and together accounting for more than 90% of total income.

R. kutum is a migratory anadromous fish endemic to the sea and spreads throughout the Caspian basin, mainly in the southern CS (Valipour and Khanipour 2006, Fazli et al. 2022a). Its spawning occurs in April (Daryanabard et al. 2020). This bottom-feeding species consumes a variety of prey items, primarily bivalves and cirripeds (Afraei Bandpei et al. 2009).

C. aurata, identified as the second important species (Fazli et al. 2022a), was introduced as an exotic species from the Black Sea between 1930 and 1934. Being euryhaline and eurythermic, C. aurata feeds on periphyton, detritus and invertebrates found throughout the CS. During spring, it tends to migrate to the northern regions, while in autumn, it moves towards the southern regions (Probatov and Tereshchenko 1951, Fazli et al. 2013). Unlike R. kutum, C. aurata is catadromous and spawns mainly in October and November (Fazli et al. 2008). During the past two decades, both C. aurata and C. saliens have experienced severe disease outbreaks, resulting in high mortality rates. A nodavirus-like agent has been isolated from both species, exhibiting clinical signs such as severe abdominal distention, abnormal swimming and hyperinflation of the swim bladder (Zorriehzahra et al. 2005, Soltani et al. 2010, Ghiasi et al. 2016).

Time series studies of population biological data for aquatic species are critical for managing the related ecosystems. These studies are valuable in several ways, including the fact that they can help establish baselines and provide insights into the patterns and factors that influence population patterns and their effects. Understanding how environmental variability and anthropogenic stressors affect long-term population dynamics will enable us to manage better and conserve ecosystems and species. Long-term changes in populations can be comprehended by employing multivariate time series analyses. Dynamic Factor Analysis (DFA) is a valuable tool for examining ecological time series data and identifying common population trends. DFA can also be utilized to evaluate the covariance among multiple ecological time series (Zuur et al. 2003). Furthermore, numerous studies have utilized DFA to assess the influence of environmental variables on ecological indices (Jorgensen et al. 2016, Barber et al. 2019, Bitetto et al. 2019). In a recent study in the southern CS, DFA was used to determine the impact of environmental variables on sturgeon populations (Fazli et al. 2022b), but DFA has never been applied to the two main Caspian benthopelagic fish stocks (C. aurata and R. kutum) in this ecosystem.

The objective of the present study was to analyse the trends of biological indicators for C. aurata and R. kutum in the southern CS. We used DFA to assess the relative synchrony of total catch and abundance, spawning stock biomass (SSB), percentage of immature and mega spawners, and relative condition factor. Also, for R. kutum the seventh response variable “fingerling released” was added to the models. The aims of the study were i) to analyse the trends of biological indicators, ii) to investigate the influence of environmental factors [SSL, seawater surface temperature (SST), the East Atlantic-West Russian pattern and river discharges], and anthropogenic activity (fishing effort) on the variability of C. aurata and R. kutum bio-indicators over the past three decades. These findings will help to better understand the Caspian ecosystem and aid in the formulation of effective fisheries management strategies.

The benthopelagic species (C. aurata and R. kutum) feed mainly on detritus and macrobenthos, and they could be impacted by the changes in the quality and quantity of the CS benthic community. Because benthic time series are rare in the study area, we used different environmental variables that affect the production and structure of both benthic and benthopelagic fish.

Many factors influence the condition of ecosystems in the CS. In the last century, the deterioration in the environmental conditions in the sea can be attributed primarily to changes in SSL and the volume of water entering local rivers, as well as pollution from a wide range of sources such as petroleum hydrocarbons and phenols (Kosarev 2005). Based on our experience and previous studies, since the 1990s three significant challenges have emerged as major threats to this ecosystem: illegal fishing and overfishing, exotic species invasion and global warming. These notable anomalies pose a direct threat because they lead to spatiotemporal gradients in the abundance of target species and habitat destruction. Additionally, they pose an indirect threat by affecting community structure and producing differential effects on functional groups. As a result, these challenges pose significant risks to the structure and function of the Caspian ecosystem (Pourang et al. 2016, Shiganova et al. 2023).

Given these challenges and their impact on the ecosystem, it is critical to examine the population dynamics of key species alongside environmental variables to gain insights into ecosystem health. In this study, we estimated six population indices, including Cat, Abu, SSB, Imm, Meg and Kn of two valuable benthopelagic species (and one more index of FR for R. kutum) from 1991 to 2020, which are useful for defining changes in their populations. Through DFA analysis, we examined common trends in the biological indices of each species. This approach allowed us to identify common trends and investigate their correlations with environmental indices. The results strongly supported the hypothesis that most biological indices of both species exhibit interannual patterns influenced by environmental variables. Notably, the patterns observed in the stock condition indices of the two benthopelagic fish displayed distinct trends over the past 30 years, which could potentially be linked to regional and local climate indices. The selected models identified one common trend with the exploratory covariates of SST, SSL and E for C. aurata and two common trends with the exploratory covariates of SST and SSL for R. kutum in the datasets spanning 30 years from 1991 (Fig. 7). Also, according to Zuur et al. (2003), if the covariates were not selected in the best models, it does not mean that these covariates did not affect the time series. The current study showed a strong negative correlation between Kn and SST and a strong positive correlation between Kn and SSL of both species.

The increasing trend of SST has been reported in several global studies conducted over the past decades (Barber et al. 2019, Bitetto et al. 2019, Feuilloley et al. 2020), which could be attributed to climate change (Xu et al. 2021). The increases in SST contribute to other environmental changes: noteworthy examples include the recruitment of R. kutum and C. aurata in response to the changes in SST in the late 1990s (Beyraghdar Kashkooli et al. 2017). Additionally, the declining recruitment and SSB of two small pelagic fish species, Clupeonella engrauliformis and C. grimmi, in the 2000s were influenced by SST and SSL (Fazli et al. 2021).

Kn serves as an indicator of environmental suitability, providing insight into the overall well-being and feeding activity of fish within a water body. A decrease in Kn below 1 signifies a decline in condition and feeding activity (Bagenal and Tesch 1978, Anderson and Neumann 1996, Sutton et al. 2000). According to the present study, the Kn of both species has shown a decreasing trend over the past three decades. However, only the Kn of C. aurata has declined, falling below 1 since the 2010s (Fig. 3), which coincides with a reduction in SSL. On the other hand, about 60% of the surface area of the CS is covered by its shelf zone. Therefore, the decrease in SSL has affected a wide area of the shelf zone, especially the North Caspian (covered by shallow waters), and can have a strong impact on the nutrition of the C. aurata that migrate to the North Caspian for feeding.

According to Ghasemi et al. (2014), following the invasion of Mnemiopsis leidyi into the Caspian ecosystem in the late 1990s, the composition of Caspian macrobenthos has undergone significant changes. Under the new conditions, numerous endemic macrobenthic species were displaced by the non-native species Streblospio gynobranchiata in the 2000s. As a result, the benthopelagic fish species, including C. aurata and R. kutum, have experienced substantial alterations in their diets over the past two decades, leading to a reduction in their Kn.

Meg is an indicator of population structure where there is no upper size limit for captures (in the case of both species in this study), with 30% to 40% of Meg representing a healthy population structure (Froese 2004). In the present study, the Meg of C. aurata was more than about 20% in 1994-2010, dropping to around 15% in the 2010s, and the Meg of R. kutum was less than 13%. Consequently, based on the Meg indicator, both species show an unhealthy population structure. Froese (2004) suggested that all of the harvested fish should be from mature fishes (Imm should be 0%) as a target. Based on this study, approximately 12% of C. aurata and 45% of R. kutum were immature fish, indicating that they were caught before they could fully realize their growth potential. This suggests that growth overfishing may have occurred in both species, especially R. kutum, over the last three decades.

E can have various effects on the Cat and Abu of fish species. The relationship between Cat and E and measures of Abu can be influenced by factors such as target fish species, fishing methods and management strategies (Mulazzani et al. 2015). In the CS, the increase in E until 2001 led to an increase in Cat of C. aurata until 2002, which caused overfishing, and despite the decrease in E in the following years, the Cat and Abu continued to decrease. In the case of R. kutum, the decrease in catch and abundance occurred three years later (in 2005).

Among the environmental (SST and SSL) and anthropogenic (E) variables affecting the biological indicators of C. aurata and R. kutum, only E can be controlled. The process of reducing E can prevent overfishing and increase the population. The effects of environmental variables on fish in the CS appear to be profound. Therefore, despite the decreasing trend of E, the Abu, Cat and Kn of the two species are still decreasing. Due to the increasing trend of global temperature and the decreasing SSL of the CS, it is likely that the decreasing trend of the biological indicators, especially the Cat, Abu, SSB and Kn of the two mentioned species, will continue in the years to come.

This study has provided a comprehensive analysis of the biological indicators of two commercially valuable benthopelagic species, C. aurata and R. kutum, within the CS over 30 years. By employing DFA, we discerned patterns in biological indices and their correlations with environmental and anthropogenic variables. We observed that SST, SSL and E have played a notable role in shaping the life history traits of C. aurata and R. kutum, as evidenced by the common trends identified. Also, the strong negative correlation between Kn and SST and the positive correlation between Kn and SSL suggest that rising temperatures may continue to adversely affect the health and vitality of these species. Furthermore, our study has highlighted the detrimental impact of overfishing, which is evident in the significant proportion of immature fish in catches and the downward trend in population metrics such as Cat, Abu and Kn.