Past Inputs Research Articles

Integrating legacy soil phosphorus into sustainable nutrient management strategies for future food, bioenergy and water security

Legacy phosphorus (P) that has accumulated in soils from past inputs of fertilizers and manures is a large secondary global source of P that could substitute manufactured fertilizers, help preserve critical reserves of finite phosphate rock to ensure future food and bioenergy supply, and gradually improve water quality. We explore the issues and management options to better utilize legacy soil P and conclude that it represents a valuable and largely accessible P resource. The future value and period over which legacy soil P can be accessed depends on the amount present and its distribution, its availability to crops and rates of drawdown determined by the cropping system. Full exploitation of legacy P requires a transition to a more holistic system approach to nutrient management based on technological advances in precision farming, plant breeding and microbial engineering together with a greater reliance on recovered and recycled P. We propose the term ‘agro-engineering’ to encompass this integrated approach. Smaller targeted applications of fertilizer P may still be needed to optimize crop yields where legacy soil P cannot fully meet crop demands. Farm profitability margins, the need to recycle animal manures and the extent of local eutrophication problems will dictate when, where and how quickly legacy P is best exploited. Based on our analysis, we outline the stages and drivers in a transition to the full utilization of legacy soil P as part of more sustainable regional and global nutrient management.

Long-Term Persistence of an Anxiolytic Drug (Oxazepam) in a Large Freshwater Lake.

Production and human consumption of pharmaceuticals result in contamination of surface waters worldwide. Little is known about the long-term (i.e., over decades) fate of pharmaceuticals in aquatic systems. Here, we show that the most prescribed anxiolytic in Sweden (oxazepam) persists in its therapeutic form for several decades after being deposited in a large freshwater lake. By comparing sediment cores collected in 1995 and 2013, we demonstrate that oxazepam inputs from the early 1970s remained in the sediments until sampling in 2013, despite in situ degradation processes and sediment diagenesis. In laboratory and pond experiments, we further reveal that therapeutic forms of oxazepam can persist over several months in cold (5 °C) lake water free from UV light. We conclude that oxazepam can persist in lakes over a time scale much longer than previously realized and that levels can build up in lakes due to both a legacy of past inputs and a growing urban population.

Combining benthic foraminiferal ecology and shell Mn/Ca to deconvolve past bottom water oxygenation and paleoproductivity

The Mn/Ca of carbonate tests of living deep-sea foraminifera (Hoeglundina elegans, Bulimina aculeata, Uvigerina peregrina and Melonis barleeanus) were determined together with pore water manganese along a bottom water oxygen gradient across the lower boundary of the Arabian Sea oxygen minimum zone. Although Mn has long been considered an indicator for contamination, new cleaning protocols and high-resolution laser ablation ICP-MS now allow the reliable analyses of test-associated Mn. Within locations, Mn incorporation between species varies as a function of their in-sediment depth preferences and associated pore water chemistry. Under well-oxygenated bottom water conditions, shallow infaunal species incorporate little Mn in their test, whereas the species collected from deeper habitats show elevated Mn concentrations. With decreasing oxygen contents pore water Mn concentrations and benthic foraminiferal in-sediment distribution change. Whereas Mn/Ca in shallow infaunal species responds moderately to bottom water oxygenation, Mn/Ca of the infaunal species M. barleeanus correlates well to oxygenation. Although high productivity results in a shallower redox cline within the sediment, pore water Mn is retained as long as the bottom water remains oxygenated. Under reduced bottom water oxygen conditions, Mn escapes to the overlying water column and test-associated Mn/Ca decreases also in the infaunal species. By combining pore water chemistry of Mn, calcitic Mn/Ca and foraminiferal ecology, a new conceptual model is presented (TROXCHEM3) that provides a framework for deconvolving past organic matter input and bottom water oxygenation.

Holocene environmental change at the southern coast of Korea based on organic carbon isotope (δ13C) and C/S ratios

To trace the environmental change that occurred at a coastal area during the Holocene, we performed radiocarbon dating and geochemical analyses using sedimentary cores recovered from the Beolgyo tidal flats on the south coast of Korea. The δ13C values of total organic carbon (TOC), which represent relative inputs of terrestrial (or freshwater) and marine organic sources, revealed a long-term increasing trend from −27.5 to −21.7‰ during the period between 10,000 and 3000 cal BP, suggesting an evolutionary series of fluvial-dominant to marine-dominant environments, coupled with weakened summer monsoon intensity. From 3000 cal BP to the present, the δ13CTOC decreased gradually from −21.7 to −24‰, indicating enhancement of freshwater input during this time. The observed long-term environmental change was divided into six stages based on the averages of the total sulfur content (TS%) and the organic carbon to sulfur ratio (C/S), proxies for paleosalinity. On multi-centennial to millennial timescales, higher C/S ratios and lower TS% were linked to higher δ13CTOC values, indicating strong freshwater input at Stage 2 (∼8300–7400 cal BP), Stage 4 (∼4500–3300 cal BP), and Stage 6 (∼1100 cal BP to present). Strong freshwater discharge events synchronously occurred at 4300 cal BP and 3600 cal BP along the south coast of Korea. Further testing should be carried out in other East Asian coastal areas. This study suggests that sedimentary δ13CTOC values and C/S ratios can be used in Asian coastal regions to trace past freshwater input events and summer monsoon intensity during relatively stable sea level periods (e.g., the middle to late Holocene).

Design of memoryless output feedback controller of discrete‐time systems with input delay

This study addresses the output feedback stabilisation problem of discrete-time linear systems with input delay. By viewing the difference between the current input and the delayed input as a special disturbance, full- and reduced-order extended state observers are constructed, respectively, to estimate both the state and the disturbance simultaneously. Then the composite control laws are synthesised to actively compensate the input delay by using the state and disturbance estimates. Different from existing predictor-based control approaches, the proposed controllers are memoryless, and the past input information are thus not needed. Moreover, it can be observed that the information on the size of the input delay is not involved in either the observers or the controllers. Finally, simulation results are provided to illustrate the advantages and effectiveness of the proposed approaches.

Nonlinear adaptive filtering using kernel‐based algorithms with dictionary adaptation

SummaryNonlinear adaptive filtering has been extensively studied in the literature, using, for example, Volterra filters or neural networks. Recently, kernel methods have been offering an interesting alternative because they provide a simple extension of linear algorithms to the nonlinear case. The main drawback of online system identification with kernel methods is that the filter complexity increases with time, a limitation resulting from the representer theorem, which states that all past input vectors are required. To overcome this drawback, a particular subset of these input vectors (called dictionary) must be selected to ensure complexity control and good performance. Up to now, all authors considered that, after being introduced into the dictionary, elements stay unchanged even if, because of nonstationarity, they become useless to predict the system output. The objective of this paper is to present an adaptation scheme of dictionary elements, which are considered here as adjustable model parameters, by deriving a gradient‐based method under collinearity constraints. The main interest is to ensure a better tracking performance. To evaluate our approach, dictionary adaptation is introduced into three well‐known kernel‐based adaptive algorithms: kernel recursive least squares, kernel normalized least mean squares, and kernel affine projection. The performance is evaluated on nonlinear adaptive filtering of simulated and real data sets. As confirmed by experiments, our dictionary adaptation scheme allows either complexity reduction or a decrease of the instantaneous quadratic error, or both simultaneously. Copyright © 2015 John Wiley & Sons, Ltd.

Visual homeostatic processing in V1: when probability meets dynamics.

Visual homeostatic processing in V1: when probability meets dynamics.

Importance of vegetation for manganese cycling in temperate forested watersheds

AbstractMany surface soils are enriched in metals due to anthropogenic atmospheric inputs. To predict the persistence of these contaminants in soils, factors that impact rates of metal removal from soils into streams must be understood. Experiments at containerized seedling (mesocosm), pedon, and catchment scales were used to investigate the influence of vegetation on manganese (Mn) transport at the Susquehanna/Shale Hills Critical Zone Observatory (SSHCZO) in Pennsylvania, USA, where past atmospheric inputs from industrial sources have enriched Mn in surface soils. Large quantities of Mn that were leached from soil components into solution were taken up by vegetation; as a result, only relatively small quantities of Mn were removed from soil into effluent and streams. Manganese uptake into vegetation exceeded Mn losses in soil leachate by 20–200X at all scales, and net Mn loss from soils decreased in the presence of vegetation due to uptake into plant tissues. The majority of Mn taken up by forest vegetation at SSHCZO each year was returned to the soil in leaf litter and consequently immobilized as Mn oxides that formed during litter decomposition. Thus, plant uptake of Mn combined with rapid oxidation of Mn during litter decomposition contribute to long‐term retention. Current release rates of soluble Mn from SSHCZO soils were similar to release rates from the larger Susquehanna River Basin, indicating that the processes observed at SSHCZO may be widespread across the region. Indeed, although atmospheric deposition of Mn has declined, surface soils at SSHCZO and throughout the eastern United States remain enriched in Mn. If recycling through vegetation can attenuate the removal of Mn from soils, as observed in this study, then Mn concentrations in soils and river waters will likely decrease slowly over time following watershed contamination. Understanding the role of vegetation in regulating metal transport is important for evaluating the long‐term effects of historical and ongoing metal loading to soils.

On Zero-Rate Error Exponents of Finite-State Channels With Input-Dependent States

We derive a single-letter formula for the zero-rate reliability (error exponent) of a finite-state channel whose state variable depends deterministically (and recursively) on past channel inputs, where the code complies with a given channel input constraint. Special attention is then devoted to the important special case of the Gaussian channel with intersymbol interference, where more explicit results are obtained.

Optimal Tracking Control of Unknown Discrete-Time Linear Systems Using Input-Output Measured Data.

In this paper, an output-feedback solution to the infinite-horizon linear quadratic tracking (LQT) problem for unknown discrete-time systems is proposed. An augmented system composed of the system dynamics and the reference trajectory dynamics is constructed. The state of the augmented system is constructed from a limited number of measurements of the past input, output, and reference trajectory in the history of the augmented system. A novel Bellman equation is developed that evaluates the value function related to a fixed policy by using only the input, output, and reference trajectory data from the augmented system. By using approximate dynamic programming, a class of reinforcement learning methods, the LQT problem is solved online without requiring knowledge of the augmented system dynamics only by measuring the input, output, and reference trajectory from the augmented system. We develop both policy iteration (PI) and value iteration (VI) algorithms that converge to an optimal controller that require only measuring the input, output, and reference trajectory data. The convergence of the proposed PI and VI algorithms is shown. A simulation example is used to verify the effectiveness of the proposed control scheme.

PO-MOESP subspace identification of Directed Acyclic Graphs with unknown topology

In this paper, a PO-MOESP subspace identification algorithm for Directed Acyclic Graphs (DAG) is presented. The state of every node and the structure of the graph are assumed to be unknown. The method estimates the column space of the observability matrix of every node by projecting away the input–output data from the rest of the graph. A past input–output instrumental variable (IV) approach is adopted to deal with the noise. The topology of the network is revealed by using dedicated projection matrices. Moreover, we provide a characterization of the class of graphs whose topology can be fully reconstructed by using this method. Finally, two simulation examples are provided to show the effectiveness of the proposed methodology.

Model-based Fault-Tolerant Control of Uncertain Particulate Processes: Integrating Fault Detection, Estimation and Accommodation

This work presents a methodology for the integrated identification, estimation and accommodation of control actuator faults in particulate processes with discretely-sampled measurements and plant-model mismatch. Initially, a stabilizing state feedback controller is designed on the basis of a reduced-order model of the infinite-dimensional system, and the closed-loop stability region is characterized in terms of the model uncertainty, the fault magnitude, the sampling period and the control design parameters. When state measurements are unavailable, the reduced-order inter-sample model predictor generates state estimates which are updated at each sampling time. A moving-horizon optimization problem is then formulated and solved for on-line actuator fault detection, isolation and estimation using past state and input data. The resulting estimates are used to locate the operating point with respect to the closed-loop stability region, which in turn is used to carry out the fault accommodation logic via updating the pot-fault control model and/or adjusting the controller design parameters. The developed methodology is illustrated using a non-isothermal continuous crystallizer example.

A comparative study of neuro fuzzy and recurrent neuro fuzzy model-based controllers for real-time industrial processes

Nonlinearities in system dynamics and the multivariable nature of processes offer a stiff challenge in designing predictive controllers that improve process performance in industries. This investigation presents a recurrent neuro fuzzy network (RNFN) model for a nonlinear multivariable system in process industries and a methodology to design model-predictive controllers (MPCs) using the proposed model. The RNFN model combines the learning features of artificial neural networks with human cognition capabilities of fuzzy systems. Therefore, RNFN leads to a modelling framework that has the ability not only to learn the model parameters, but also makes decision on operating region of the nonlinear model depending on the input–output data. Furthermore, the recurrent structure and the introduction of a memory unit between the fuzzy inference and fuzzification layer enhance the prediction capability due to the use of past input–output data, making the model more suitable for designing predictive controllers. Next, the MPC design methodology that exploits the advantages of the RNFN model to optimize the control moves is presented. The proposed MPC uses the gradient descent algorithm to minimize the control moves as against the traditional state-space approaches that require complex computations and solvers. Therefore, implementing the proposed MPC in embedded hardware becomes easier. The proposed modelling framework and the MPC design methodology are illustrated using experiments on a laboratory-scale quadruple tank. Our experiments show that the proposed RNFN-based MPC performs better than the neuro fuzzy network-based MPC for both servo and regulatory responses.

Prediction of financial time series and its volatility using a hybrid dynamic neural network trained by sliding mode algorithm and differential evolution

A dynamic neural network (DNN) and a new computationally efficient functional link artificial neural network (CEFLANN) combination optimised with differential evolution (DE) is presented in this paper to predict financial time series like stock price indices and stock return volatilities of two important Indian stock markets, namely the Reliance Industries Limited (RIL), and NIFTY from one day ahead to one month in advance. The DNN comprises a set of 1st order IIR filters for processing the past inputs and their functional expansions and its weights are adjusted using a sliding mode strategy known for its fast convergence and robustness with respect to chaotic variations in the inputs. Extensive computer simulations are carried out to predict simultaneously the stock market indices and return volatilities and it is observed that the simple IIR–based DNN–FLANN model hybridised with DE produces better forecasting accuracies in comparison to the more complicated neural architectures.

Adaptive correlation filter-based video stabilization without accumulative global motion estimation

We present a digital video stabilization approach that provides both robustness and efficiency for practical applications. In this approach, we adopt a stabilization model that maintains spatio-temporal information of past input frames efficiently and can track original stabilization position. Because of the stabilization model, the proposed method does not need accumulative global motion estimation and can recover the original position even if there is a failure in interframe motion estimation. It can also intelligently overcome the situation of damaged or interrupted video sequences. Moreover, because it is simple and suitable to parallel scheme, we implement it on a commercial field programmable gate array and a graphics processing unit board with compute unified device architecture in a breeze. Experimental results show that the proposed approach is both fast and robust.

A novel control algorithm for dephosphorization in an LD converter

This paper proposes a new control algorithm for dephosphorization in the steelmaking process using a LD converter. The control algorithm is composed of two steps, which achieves a low phosphorous content in the hot metal after treatment. The first step creates target concentration for the iron oxide in the slag for the next heat, which is optimal to achieve effective dephosphorization. The target concentration is obtained after applying independent component analysis (ICA) to the concentration data of past heats. The second step decides the control inputs for the next heat, namely the oxygen blowing rate and the lance height by calculating a weighted summation of past control inputs. The weight is based on the difference between the target concentration and concentration of each past heat. The proposed algorithm was applied to an actual process, and the results showed high control performance at the end of this paper.

Identification of functional synaptic plasticity from spiking activities using nonlinear dynamical modeling

This paper presents a systems identification approach for studying the long-term synaptic plasticity using natural spiking activities. This approach consists of three modeling steps. First, a multi-input, single-output (MISO), nonlinear dynamical spiking neuron model is formulated to estimate and represent the synaptic strength in means of functional connectivity between input and output neurons. Second, this MISO model is extended to a nonstationary form to track the time-varying properties of the synaptic strength. Finally, a Volterra modeling method is used to extract the synaptic learning rule, e.g., spike-timing-dependent plasticity, for the explanation of the input–output nonstationarity as the consequence of the past input–output spiking patterns. This framework is developed to study the underlying mechanisms of learning and memory formation in behaving animals, and may serve as the computational basis for building the next-generation adaptive cortical prostheses.

Nd and Sr isotope compositions of different phases of surface sediments in the South Pacific: Extraction of seawater signatures, boundary exchange, and detrital/dust provenance

The radiogenic isotope composition of neodymium (Nd) and strontium (Sr) are useful tools to investigate present and past oceanic circulation or input of terrigenous material. We present Nd and Sr isotope compositions extracted from different sedimentary phases, including early diagenetic Fe-Mn coatings, “unclean” foraminiferal shells, fossil fish teeth, and detritus of marine surface sediments (core-tops) covering the entire midlatitude South Pacific. Comparison of detrital Nd isotope compositions to deep water values from the same locations suggests that “boundary exchange” has little influence on the Nd isotope composition of western South Pacific seawater. Concentrations of Rare Earth Elements (REE) and Al/Ca ratios of “unclean” planktonic foraminifera suggest that this phase is a reliable recorder of seawater Nd isotope composition. The signatures obtained from fish teeth and “nondecarbonated” leachates of bulk sediment Fe-Mn oxyhydroxide coatings also agree with “unclean” foraminifera. Direct comparison of Nd isotope compositions extracted using these methods with seawater Nd isotope compositions is complicated by the low accumulation rates yielding radiocarbon ages of up to 24 kyr, thus mixing the signal of different ocean circulation modes. This suggests that different past seawater Nd isotope compositions have been integrated in authigenic sediments from regions with low sedimentation rates. Combined detrital Nd and Sr isotope signatures indicate a dominant role of the Westerly winds transporting lithogenic material from South New Zealand and Southeastern Australia to the open South Pacific. The proportion of this material decreases toward the east, where supply from the Andes increases and contributions from Antarctica cannot be ruled out.

A Sliding-Window Modeling Approach for Neural Network

A sliding-window modeling approach of neural network (SWMANN) was presented. The basic idea is that training data for neural network should be reconstructed by means of a slide-window way to build input and target samples. It means that the output is determined not only by the current input, but also by the past input, which could better follow the dynamic change and development of real systems. In this paper, SWMANN was introduced. A detailed theoretical derivation was described for its implementation by taking wavelet neural network as a modeling tool. Moreover, SWMANN was compared with classical modeling approach of neural network (CMANN). Theoretical results indicated that CMANN is a special case of SWMANN when sliding-window parameters are selected as 1. Therefore, compared with CMANN, SWMANN presented in this paper is a more general form.

Estimation of past and recent carbon input by crops into agricultural soils of southeast Germany

In agricultural soils, the formation of soil organic matter largely depends on the carbon (C) input by crop residues and rhizodeposition, which is thus of decisive importance for the management and prediction of soil organic carbon (SOC) stocks in cropland and grassland. However, there is a remarkable lack of reliable, crop-specific C input data. We used a plant C allocation approach to estimate the C input of major crops and grassland into agricultural soils of Bavaria in southeast Germany. Historic and recent plant C allocation coefficients were estimated and C inputs were calculated for a 60-year period (1951–2010) using long-term agricultural statistics. The spatial distribution of C inputs within Bavaria was derived from county-specific statistical data. The results revealed increases of the C input by 107–139% for cereals, 173–188% for root, forage and leguminous crops and 34% for grassland in the last 60 years. This increase was related to linear yield increases until 1995 despite significant changes of plant C allocation. However, from 1995 onwards, crop yields and related C inputs stagnated, which allowed a robust estimation of recent crop-specific C input values. A total C input of 3.8–6.7tha−1yr−1 was estimated for cereals, 5.2–6.3tha−1yr−1 for root, forage and leguminous crops and 2.4tha−1yr−1 for grassland. These amounts were partly higher compared to estimations in the literature. A generally high spatial variability of C inputs was detected within Bavaria with differences of up to 40% between adjacent counties. The results of this study could be used to optimize the C input of crop rotations and thus promote the formation of soil organic matter and C sequestration in agricultural soils on the basis of a soil carbon model. Moreover, recent estimations of C inputs could be used to model the future development of agricultural SOC stocks. A further stagnation of crop yields and the related C input under an ongoing temperature increase bears the risk of a future decrease of SOC stocks in cropland soils of Bavaria.