Dissolved Oxygen Time Series Research Articles

Historical climate drivers and species’ ecological niche in the Beaufort Sea food web

Abstract Climate change impacts have been particularly acute and rapid in the Arctic, raising concerns about the conservation of key ecologically and culturally significant species (e.g. beluga whales, Arctic cod), with consequences for the Indigenous community groups in the region. Here, we build on an Ecopath with Ecosim model for the Canadian Beaufort Sea Shelf and Slope to examine historical (1970–2021) changes in the ecological dynamics of the food web and key species under climate change. We compare the individual and cumulative effects of (i) increased sea surface temperature; (ii) reduced sea ice extent; (iii) ocean deoxygenation; and (iv) changing ocean salinity in the ecosystem models. We found that including salinity time series in our ecosystem models reduced the diversity found within the ecosystem, and altered the trophic levels, biomass, and consumption rates of some marine mammal and fish functional groups, including the key species: beluga whales, as well as Arctic and polar cods. Inclusion of the dissolved oxygen time series showed no difference to ecosystem indicators. The model findings reveal valuable insights into the attribution of temperature and salinity on Arctic ecosystems and highlight important factors to be considered to ensure that existing conservation measures can support climate adaptation.

A hybrid intelligence model for predicting dissolved oxygen in aquaculture water

Dissolved oxygen is an important water quality indicator that affects the health of aquatic products in aquaculture, and its monitoring and prediction are of great significance. To improve the prediction accuracy of dissolved oxygen water quality series, a hybrid prediction model based on variational mode decomposition (VMD) and a deep belief network (DBN) optimized by an improved slime mould algorithm (SMA) is proposed in this paper. First, VMD is used to decompose the nonlinear dissolved oxygen time series into several relatively stable intrinsic mode function (IMF) subsequences with different frequency scales. Then, the SMA is improved by applying elite opposition-based learning and nonlinear convergence factors to increase its population diversity and enhance its local search and global convergence capabilities. Finally, the improved SMA is used to optimize the hyperparameters of the DBN, and the aquaculture water quality prediction VMD-ISMA-DBN model is constructed. The model is used to predict each IMF subsequence, and the ISMA optimization algorithm is used to adaptively select the optimal hyperparameters of the DBN model, and the prediction results of each IMF are accumulated to obtain the final prediction result of the dissolved oxygen time series. The dissolved oxygen data of aquaculture water from 8 marine ranches in Shandong Province, China were used to verify the prediction performance of the model. Compared with the stand-alone DBN model, the prediction performance of the model has been significantly improved, MAE and MSE have been reduced by 43.28% and 40.43% respectively, and (R2) has been increased by 8.37%. The results show that the model has higher prediction accuracy than other commonly used intelligent models (ARIMA, RF, TCN, ELM, GRU and LSTM); hence, it can provide a reference for the accurate prediction and intelligent regulation of aquaculture water quality.

Assessment of multiple ecosystem metabolism methods in an estuary

AbstractEcosystem metabolism is a key ecological attribute and easy to describe, but quantifying metabolism in estuaries is challenging. Properly scaling measurements through time and space requires consideration of hydrodynamics and mixing water from heterogeneous sources, making any estimation uncertain. Here, we compared three methods for modeling ecosystem metabolism in a portion of the Sacramento‐San Joaquin Delta. Metabolism estimates based on laboratory incubations, continuous in situ buoys, and an oxygen isotope approach all indicated the system was net heterotrophic, and calculated rates were comparable in magnitude when averaged over the 2‐month study. Daily metabolic rates based on in situ buoys were the most variable, likely due to horizontal and vertical advection and poor portrayal of the dissolved oxygen budget. After temporally averaging in situ buoy estimates or smoothing the dissolved oxygen time series for tidal effects, rates were more comparable to the other methods, which may be necessary to account for tidal advection and unbalanced contributions from subhabitats within the metabolic footprint. Incubation‐based rates represent the finest temporal and spatial scale and only account for pelagic processes, which may explain why incubation‐based rates were lower than the other two methods. The oxygen isotope method provided temporally and spatially integrated rates that were bracketed by the other two methods and may be a valuable tool in systems matching the model requirements. Because uncertainty arises in each method from a number of assumptions and scaling calculations, the resolution of metabolic rates in estuaries is likely coarser and more variable than in other aquatic ecosystems.

Predicting river dissolved oxygen time series based on stand-alone models and hybrid wavelet-based models

Accurate prediction of dissolved oxygen time series is important for improving the water environment and aiding water resource management. In this study, four stand-alone models including multiple linear regression (MLR), support vector machine (SVM), artificial neural network (ANN) and random forest (RF), and four hybrid models based on wavelet transform (WT) including WT-MLR, WT-SVM, WT-ANN and WT-RF were used to predict the daily dissolved oxygen (DO) at 1–5-day lead times in the Dongjiang River Basin, China. To make the prediction robust, the maximal information coefficient (MIC) was used to capture comprehensive information between DO and explanatory variables. The 5-fold cross validation grid search approach was used to optimize parameters of machine learning tools. Two types of frameworks of WT: direct framework (i.e., only the explanatory variables were decomposed) and multicomponent framework (i.e., both explanatory variables and target variables were decomposed) were used to construct hybrid models. The results show that MIC extracts four optimal explanatory variables: previous DO, water temperature, air temperature and air pressure. Four evaluation parameters including correlation coefficient (R), Nash-Sutcliffe efficiency (NSE), mean absolute error (MAE) and root mean square error (RMSE) indicate that the prediction accuracy decreases as the lead time changes from 1 to 5 days. In terms of the stand-alone models, MLR model outperforms the other three models with higher NSE values of 0.616–0.921, and lower RMSE values of 0.503–1.111. With regard to the hybrid models, WT-ANN and WT-MLR models exhibit higher performance, and multicomponent framework performs better than direct framework in all hybrid models. In general, the multicomponent framework of WT can improve the prediction accuracy of stand-alone models at a certain degree, while the direct framework shows no obvious advantage.

Physically Based Scaling Models to Predict Gas Transfer Velocity in Streams and Rivers

AbstractAir‐water gas transfer influences fundamental processes of aquatic systems such as photosynthesis, respiration, and greenhouse gas emission. Previous predictive models for gas transfer velocity (KL) often only apply to a small range of streams and rivers. Two “universal” scalings, KL ∼ 1/4 power of turbulent dissipation rate (ε1/4) and KL ∼ shear velocity (U*), deduced from classic thin‐film theory and surface renewal theory, have been reported for lake and marine systems, but have not been adequately tested for streams or rivers. To test the two scalings and understand how KL varies across systems, we compiled 588 field measurements of KL, and used inverse modeling based on long‐term high‐frequency dissolved oxygen time series to estimate KL of 35 streams and rivers capturing a wide range of discharge (0–221 m3/s). We found that the two scalings held for both the KL estimated by inverse modeling and the KL from measurements and outperformed 23 previous models in predicting KL. Furthermore, the main driver of KL variation shifts from bottom friction to other factors with increasing stream/river size. In most streams and a few rivers, correlation between U* and KL is high, and discharge‐KL relationships are strongly positive, indicating that turbulence generated by bottom friction dominates near‐surface turbulence, a fundamental control of KL, thus, the two scaling models are applicable to these systems. However, most rivers (mean discharge >10 m3/s) show low correlations between U* and KL, suggesting biochemical factors and winds likely override bottom friction in driving near‐surface turbulence and KL.

Prediction of dissolved oxygen in pond culture water based on K-means clustering and gated recurrent unit neural network

Dissolved oxygen in water is an important ecological factor in ensuring the healthy growth of aquatic products, as hypoxic stress is known to restrict the growth of aquatic products. The accurate monitoring and prediction of dissolved oxygen is the key to precise regulation and control of pond aquaculture water quality. The current dissolved oxygen prediction model has some limitations, such as a short prediction period and inadequate prediction accuracy for actual production demands. Therefore, a prediction model of dissolved oxygen in pond culture was proposed based on K-means clustering and Gated Recurrent Unit (GRU) neural network. Firstly, the key factors affecting the changes in dissolved oxygen were selected by principal component analysis (PCA). The dissolved oxygen time series was then subjected to K-means clustering, and the dissolved oxygen prediction model was constructed using GRU. To improve the clustering effect, we enhanced the similarity calculation for the time series based on the variation of dissolved oxygen. This process combined the Euclidean distance with the dynamic time-warping distance. The proposed method can predict the dissolved oxygen content of aquaculture water over different time intervals according to the demands of real-world scenarios. The average absolute error of the 30-min interval model was 0.264, and the mean absolute percentage error was 3.5 %. Experimental results indicated that the proposed method achieves higher prediction accuracy and flexibility than the conventional approach.

Time‐varying responses of lake metabolism to light and temperature

AbstractLight is a primary driver of lake ecosystem metabolism, and the dependence of primary production on light is often quantified as a photosynthesis‐irradiance or “P‐I” curve. The parameters of the P‐I curve (e.g., the maximum primary production when light is in excess) can change through time due to a variety of biological factors (e.g., changes in biomass or community composition), which themselves are subject to external drivers (e.g., herbivory or nutrient availability). However, the relative contribution of variation in the P‐I curve to overall ecosystem metabolism is largely unknown. I developed a statistical model of ecosystem metabolism with time‐varying parameters governing the P‐I curve, while also accounting for the influence of temperature. I parameterized the model with dissolved oxygen time series spanning six summers from Lake Mývatn, a shallow eutrophic lake in northern Iceland with large temporal variability in ecosystem metabolism. All of the estimated parameters of the P‐I curve varied substantially through time. The sensitivity of primary production to light under light‐limiting conditions was particularly variable (>15‐fold) and had a compensatory relationship with ambient light levels. However, the 3.5‐fold variation in the maximum potential primary production made the largest contribution to variation in ecosystem metabolism, accounting for around 90% of the variance in net ecosystem production. Much of the variation in maximum primary production was attributable to cyanobacterial blooms, which occur in some but not all years in Mývatn. Overall, these results illustrate how changes in the P‐I curve contribute substantially to temporal variation in lake ecosystem metabolism.

Seasonal oxygen dynamics in a warm temperate estuary: effects of hydrologic variability on measurements of primary production, respiration, and net metabolism.

Seasonal responses in estuarine metabolism (primary production, respiration, and net metabolism) were examined using two complementary approaches. Total ecosystem metabolism rates were calculated from dissolved oxygen time series using Odum's open water method. Water column rates were calculated from oxygen-based bottle experiments. The study was conducted over a spring-summer season in the Pensacola Bay estuary at a shallow seagrass-dominated site and a deeper bare-bottomed site. Water column integrated gross production rates more than doubled (58.7 to 130.9 mmol O2 m-2 d-1) from spring to summer, coinciding with a sharp increase in water column chlorophyll-a, and a decrease in surface salinity. As expected, ecosystem gross production rates were consistently higher than water column rates, but showed a different spring-summer pattern, decreasing at the shoal site from 197 to 168 mmol O2 m-2 d-1 and sharply increasing at the channel site from 93.4 to 197.4 mmol O2 m-2 d-1. The consistency among approaches was evaluated by calculating residual metabolism rates (ecosystem - water column). At the shoal site, residual gross production rates decreased from spring to summer from 176.8 to 99.1 mmol O2 m-2 d-1, but were generally consistent with expectations for seagrass environments, indicating that the open water method captured both water column and benthic processes. However, at the channel site, where benthic production was strongly light-limited, residual gross production varied from 15.7 mmol O2 m-2 d-1 in spring to 86.7 mmol O2 m-2 d-1 in summer. The summer rates were much higher than could be realistically attributed to benthic processes, and likely reflected a violation of the open water method due to water column stratification. While the use of sensors for estimating complex ecosystem processes holds promise for coastal monitoring programs, careful attention to the sampling design, and to the underlying assumptions of the methods, is critical for correctly interpreting the results. This study demonstrated how using a combination of approaches yielded a fuller understanding of the ecosystem response to hydrologic and seasonal variability.

Estimating time‐variable aerobic respiration in the streambed by combining electrical conductivity and dissolved oxygen time series

AbstractAerobic respiration is an important component of in‐stream metabolism. The larger part occurs in the streambed, where it is difficult to directly determine actual respiration rates. Existing methods for determining respiration are based on indirect estimates from whole‐stream metabolism or provide time invariant results estimated from oxygen consumption measurements in enclosed chambers that do not account for the influence of hydrological changes. In this study we demonstrate a simple method for determining time‐variable hyporheic respiration. We use a windowed cross‐correlation approach for deriving time‐variable travel times from the naturally changing electrical conductivity signal that is transferred into the sediment. By combining the results with continuous in situ dissolved oxygen measurements, variable oxygen consumption rate coefficients in the streambed are obtained. An empirical temperature relationship is derived and used for standardizing the respiration rate coefficients to isothermal conditions. For demonstrating the method, we compare two independent measurement spots in the streambed, which were located upstream and downstream of an in‐stream gravel bar and thus exposed strongly diverse travel times. The derived respiration rate results are in accordance with findings of other stream studies. By comparing the travel time and respiration rate coefficient (i.e., Damköhler number) we estimate the contribution of each to the oxygen consumption in the streambed.

Wavelet Characteristics of Hydrological and Dissolved Oxygen Time Series in a Lowland River

In this study, we investigated the temporal variability of dissolved oxygen and water temperature in conjunction with water level fluctuations and river discharge in the Narew lowland river reach. For this purpose, high resolution hydrologic and water quality time series have been used. Spectral analyses of time series using continuous wavelet transform scheme have been applied in order to identify characteristic scales, its duration, and localisation in time. The results of wavelet analysis have shown a great number of periodicities in time series at the inter-annual time scale when compared to the classical Fourier analysis. Additionally, wavelet coherence revealed the complex nature of the relationship between dissolved oxygen and hydrological variables dependent on the scale and localisation in time. Hence, the results presented in this paper may provide an alternative representation to a frequency analysis of time series.

Improving estimates of ecosystem metabolism by reducing effects of tidal advection on dissolved oxygen time series

AbstractIn aquatic systems, time series of dissolved oxygen (DO) have been used to compute estimates of ecosystem metabolism. Central to this open‐water method is the assumption that the DO time series is a Lagrangian specification of the flow field. However, most DO time series are collected at fixed locations, such that changes in DO are assumed to reflect metabolism and that effects of advection or mixing are negligible. A weighted regression model was applied to remove variability in DO time series from tides, thereby helping to partially relax this assumption and improve metabolism estimates. The method offers a distinct advantage over traditional deconvulution methods by targeting the periodicity of the tidal component while preserving the true biological signal. The model was first applied to simulated DO time series with specified biological and physical characteristics, and then applied to 1 yr of continuous monitoring data from four stations within the National Estuarine Research Reserve System. The correlation of DO and metabolism estimates with tides was greatly reduced after using weighted regression. The model was especially effective when the magnitude of tidal influence was high and correlations between tidal change and the solar cycle were low at the time scales of interest. The model was less robust when tides and the solar cycle were correlated for protracted periods. By reducing the effects of physical transport on metabolism estimates, there may be increased potential to empirically relate metabolic rates to causal factors on timescales of several days to several weeks.

Time dependent intrinsic correlation analysis of temperature and dissolved oxygen time series using empirical mode decomposition

In the marine environment, many fields have fluctuations over a large range of different spatial and temporal scales. These quantities can be nonlinear and non-stationary, and often interact with each other. A good method to study the multiple scale dynamics of such time series, and their correlations, is needed. In this paper an application of an empirical mode decomposition based time dependent intrinsic correlation, of two coastal oceanic time series, temperature and dissolved oxygen (saturation percentage) is presented. The two time series are recorded every 20min for 7years, from 2004 to 2011. The application of the empirical mode decomposition on such time series is illustrated, and the power spectra of the time series are estimated using the Hilbert transform (Hilbert spectral analysis). Power-law regimes are found with slopes of 1.33 for dissolved oxygen and 1.68 for temperature at high frequencies (between 1.2 and 12h) with both close to 1.9 for lower frequencies (time scales from 2 to 100days). Moreover, the time evolution and scale dependence of cross correlations between both series are considered. The trends are perfectly anti-correlated. The modes of mean year 3 and 1year have also negative correlation, whereas higher frequency modes have a much smaller correlation. The estimation of time-dependent intrinsic correlations helps to show patterns of correlations at different scales, for different modes.

Modelling dissolved oxygen and benthic algae dynamics in a coastal ecosystem by exploiting real-time monitoring data

In this work we propose a methodological approach for the detection and simulation of relevant changes in coastal ecosystems, i.e. oxygen depletion and proliferation of benthic algae. This approach is based on the integration of the data provided by real-time monitoring systems with the output of complex ecosystem models. We tested the method in a case study, where real-time Dissolved Oxygen (DO) data and a 2D Reaction-Transport model were used to simulate the growth of macroalgae and the daily dynamics of DO in the Lagoon of Venice (Italy). The spatiotemporal relationships among the macroalgae distribution and the DO observations were quantified by analysing and comparing the Dissolved Oxygen time series and model outputs. The outcomes were used for the inverse estimation of the initial, i.e. late winter, biomass of macroalgae. The model was then applied to simulate the growth of macroalgae and the daily dynamic of DO during the productive (i.e., spring and summer) seasons. The comparison between the model output and the real-time data indicates that the model had skill in simulating the short term (daily) DO dynamic at several lagoon sites impacted by macroalgae proliferation. The simulated intra-daily variability of Dissolved Oxygen is significantly correlated with the observations in half of the monitoring sites, as well the simulated algal growth, which resulted comparable with the field measurements. The estimated average value of macroalgae biomass was ∼1 kgfw m−2 in 2007, which is consistent with a good to moderate quality status of the Lagoon of Venice. The proposed methodology can be useful in the assessment of the environmental status of coastal ecosystems as required by recent national and international legislation.

Fuzzy logic modeling of the dissolved oxygen fluctuations in Golden Horn

The modern modeling techniques show a significant progress in recent years. Fuzzy logic approach is one of these methods that can be used for forecasting purposes and identification of complex systems. In this study, it is aimed to model monthly dissolved oxygen (DO) amount variations. Firstly, regression technique is used to remove the trend from the actual dissolved oxygen time series. Detrended time series is then modeled with Takagi–Sugeno Fuzzy logic approach. The monthly historical records of DO as considered in this paper provide a fundamental data exhibiting persistence. It is this dependence that gives opportunity for the serial modeling of the data sequence concerned. These models are the prime and sole means to derive the future likely excedences over the historical values and hence provide a chance for the future planner with decision variables such as the minimum DO amounts and their persistence for some duration. In the scope of this study, DO concentration changes in the two stations located in Golden Horn at the 0.5 m (upper layer) depth were modeled. As a result of the study, it is seen that one can forecast the next month's DO amount from antecedent measurements within an acceptable relative error limits.

Effects of small scale fluid motion on bacterial growth and respiration

Summary1. Laboratory experiments were conducted to investigate the effect of small‐scale turbulent motion on the growth and respiration of bacteria in an oscillating grid apparatus. The experiments were performed under a range of energy dissipation levels similar to those occurring in freshwater systems.2. The results showed that small‐scale turbulent motion does have an effect on bacterial growth and respiration. A higher gradient in the dissolved oxygen time series, higher 5‐day biochemical oxygen demand values, increased bacterial abundance, increased bacterial specific respiration, higher bacterial growth rate and increased nutrient uptake were all observed when the energy dissipation rate in the water column was increased.3. This has implications for traditional laboratory procedures that are used to characterise bacterial metabolic rates under stagnant fluid‐flow conditions, such as biochemical oxygen demand (BOD), which would be influenced by the effects of the small‐scale fluid motion inherent in aquatic environments. According to our results, BOD values in natural systems experiencing fluid motion would be higher than traditional bottle‐derived rates.