Niche Differentiation Characteristics of Phytoplankton Functional Groups in Arid Regions of Northwest China Based on Machine Learning

TL;DR

This study uses machine learning to analyze how phytoplankton communities in arid region reservoirs are shaped by environmental factors and interspecies interactions.

Contribution

The study reveals how environmental variables and interspecific associations influence phytoplankton stability in arid reservoirs using machine learning.

Findings

Eight dominant phytoplankton functional groups were identified, with pH and electrical conductivity as key drivers.

Positive associations in most reservoirs indicate stable communities, while SY Reservoir shows negative associations due to human disturbance.

Functional groups like S1 promote growth by reducing resource competition, while H1 inhibits others in eutrophic conditions.

Abstract

To reveal the distribution patterns of phytoplankton in reservoir ecosystems, this study selected four representative reservoirs in the Tarim River Basin, Xinjiang, China. It systematically analysed the structural characteristics of phytoplankton communities and simultaneously clarified the correlation between these communities and various environmental factors and their interrelationships. From 2023 to 2024, seasonal sampling of phytoplankton functional groups was conducted in 4 reservoirs in the upper Tarim River Basin. Eighteen functional groups were identified (8 dominant ones), with pH and electrical conductivity as key driving factors. According to the results of Redundancy Analysis (RDA), on the RDA1 axis, pH value, electrical conductivity (COND), and total dissolved solids (TDS) showed positive correlations, while all other environmental factors exhibited negative correlations; on the RDA2 axis, dissolved oxygen (DO) and pH value presented positive correlations, and all other environmental factors displayed negative correlations. According to the analysis of interspecific association using the Association Coefficient (AC), among all the reservoirs studied, only Reservoir Shangyou Reservoir (SY) exhibited a higher number of negative correlations than positive correlations in the interrelationships of phytoplankton functional groups; the other reservoirs all showed a greater number of positive correlations than negative ones. Communities in three reservoirs, including Duolang Reservoir (DL), are characterised by positive associations (stable), whereas SY is characterised by negative associations (early successional stage). This reveals interspecific interactions and mechanisms, providing a basis for ecological management.

This study investigates the distribution patterns, interspecific relationships, and community stability mechanisms of phytoplankton functional groups, aiming to elucidate the ecological processes that drive phytoplankton communities in aquatic ecosystems of arid regions. We conducted seasonal sampling from 2023 to 2024 at four auxiliary reservoirs in the Tarim River Basin, namely Shangyou Reservoir (SY), Shengli Reservoir (SL), Duolang Reservoir (DL), and Xinjingzi Reservoir (XJZ). In recent years, researchers have grouped phytoplankton into functional groups based on their shared morphological, physiological, and ecological characteristics—with these three types of traits serving as the core criteria for distinguishing different functional groups. A total of 18 functional groups were identified from the phytoplankton collected across four seasons, among which eight (A, D, H1, L0, M, MP, P, and S1) are dominant. Redundancy Analysis (RDA) indicated that environmental factors such as pH, electrical conductivity (COND), and dissolved oxygen (DO) are key driving factors affecting phytoplankton functional groups. Interspecific association analysis showed that the phytoplankton communities in DL, SL, and XJZ reservoirs were dominated by positive associations, reflecting stable community structures that are less prone to drastic fluctuations under stable environmental conditions. In contrast, the SY Reservoir was dominated by negative associations, indicating that it is in the early stage of succession with an unstable community. This may be related to intense human disturbance to the reservoir and its role in replenishing the Tarim River, which leads to significant water level fluctuations. The results of the Chi-square test and Pearson correlation analysis showed consistent trends but also differences: constrained by the requirement for continuous normal distribution, Pearson correlation analysis identified more pairs of negative associations, reflecting its limitations in analysing clumped-distributed species. Random forest models further indicated that functional groups M, MP, L0, and S1 are the main positive drivers of interspecific relationships. Among them, the increase in S1 can promote the growth of functional groups dominated by Navicula sp. and Chroococcus sp. by reducing resource competition. Conversely, the expansion of functional group H1 inhibits other groups, which is related to its adaptive strategy of resisting photo-oxidation in eutrophic environments. This study reveals the patterns of interspecific interactions and stability mechanisms of phytoplankton functional groups in arid-region reservoirs, providing a scientific basis for the management and conservation of aquatic ecosystems in similar extreme environments.

Figures