Human impacts alter driver–response relationships in lakes of Southwest China

Abstract

Biodiversity and ecological stability are closely linked, and over recent timescales, anthropogenic impacts have accelerated losses in both from local to global scales. We attempt to show the combined response of diversity and stability of an aquatic community to changes in human activity as a driver. To address this, we measured the diversity and variability of chironomids and their drivers and nature of response to external conditions over the last century, based on 4 lake sediment sequences from Southwest China, one of world's 36 biodiversity hotspots. Our results showed that the driver–response relationship was linear in a lake without direct human impacts but nonlinear in human directly impacted lakes. Recent decreases in alpha diversity and increases in beta diversity were commonly recorded in all four lakes, suggesting that both species loss and a faster replacement of chironomid taxa are a regional phenomenon. However, in the same context of human‐induced global warming, increased variability and regime shifts only occurred in lowland lakes, directly disturbed by humans, highlighting that direct human impacts have overcome natural forcing as the determinant driver shaping the chironomid composition in these sites. In addition, we found that increases in beta diversity occurred prior to a regime shift and its character depends on how the community responds to the key external pressure. Our findings reveal that direct human disturbances have largely reshaped the chironomid composition and induced an earlier regime shift at the cost of species loss, resilience loss, and a change in driver–response type.

Review on the Change Trend, Attribution Analysis, Retrieval, Simulation, and Prediction of Lake Surface Water Temperature

“…It controls the transmission by designing "gated states" through special memory units that remember important feature information and forget unimportant ones. This special design can be well done for simulation and prediction of longterm time series data [11], [240], [241]. If all influences are fed into the LSTM model, the neural network will generate a lot of noise, which will interfere with the efficiency of the model in learning useful information.…”

“…If all influences are fed into the LSTM model, the neural network will generate a lot of noise, which will interfere with the efficiency of the model in learning useful information. The possibility of model overfitting can be reduced by using principal component analysis (PCA) for the variables selected by the correlation coefficient [11].…”

“…By fitting the LST retrieved by satellite RS with the LST daily variation model in the all-day cloud-free condition, the LST data can be simulated to obtain the LST data at any time of the day, thus achieving the time normalization of the surface temperature of polar orbiting satellites in the all-day cloud-free condition [265]- [267]. In order to apply the daily variation model to surface temperature data at different observation times, Huang et al (2014) developed a series of LST daily variation models with different number of parameters (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12). This method can only be applied to cloud-free conditions throughout the day, and it cannot be used in partially cloudy conditions (i.e., no clouds during the transit time of polar-orbiting satellites) [268], found an approximately linear variation of LST with time during the morning hours through statistical analysis of the data, and used this feature to develop a time-normalized model of LST by the slope function method [269].…”

“…Longterm climate change and extreme climate events pose different levels of threat to the ecosystems of lakes. In particular, the accelerated warming of average temperatures and the frequency of extreme climate events in recent years have resulted in the ecosystems of the lakes being vulnerable: accelerated loss of biodiversity and ecological collapse [11]- [13].…”

Nonlinear dynamics, resilience, and regime shifts in aquatic communities and ecosystems: an overview

“…The future of aquatic ecosystems is contingent on past and present conditions, which makes empirical analysis challenging due to the short-term nature of most research programs. Several contributions consider historical attributes using paleo-analyses to characterize changes in communities of diatoms in Ecuadorian lakes (Benito et al 2022), and chironomids in both Chinese (Zheng et al 2022) and Alaskan (Mayfield et al 2022) lakes. While chironomid shifts show the strong impact of human disturbances, diatom communities demonstrate remarkable resilience over the past two centuries.…”