Search

Introduction: Mounting evidence suggests that geographic ranges of tree species worldwide are shifting under global environmental changes. Little is known, however, about if and how these species' range shifts may trigger the range shifts of various types of forests. Markowitz's portfolio theory of investment and its broad application in ecology suggest that the range shift of a forest type could differ substantially from the range shifts of its constituent tree species. Methods: Here, we tested this hypothesis by comparing the range shifts of forest types and the mean of their constituent species between 1970-1999 and 20002019 across Alaska, Canada, and the contiguous United States using continent- wide forest inventory data. We first identified forest types in each period using autoencoder neural networks and K-means cluster analysis. For each of the 43 forest types that were identified in both periods, we systematically compared historical range shifts of the forest type and the mean of its constituent tree species based on the geographic centroids of interpolated distribution maps. Results: We found that forest types shifted at 86.5 km-decade-1 on average, more than three times as fast as the average of constituent tree species (28.8 km-decade-1). We showed that a predominantly positive covariance of the species range and the change of species relative abundance triggers this marked difference. Discussion: Our findings provide an important scientific basis for adaptive forest management and conservation, which primarily depend on individual species assessment, in mitigating the impacts of rapid forest transformation under climate change. [ABSTRACT FROM AUTHOR]

In the present work, existing empirical expressions for longitudinal dispersion coefficient of rivers (K) are evaluated. They are found inadequate primarily because these expressions ignore the channel sinuosity, an important parameter representing a river’s transverse irregularities that affect mixing process. Hence, a new expression for K is derived taking into account the sinuosity besides few of other hydraulic and geometric characteristics of a river. The model makes use of genetic algorithm (GA) on published field data. Based on several performance indices, the new expression is found superior to many existing expressions for estimating K. The sensitivity and error analysis conducted on parameters of the new expression show the channel sinuosity an important input for predicting K accurately. Any error in measurement of sinuosity would lead to significant deviation in the longitudinal dispersion coefficient in sinuous rivers.

The reliable detection and neutralization of unmanned aircraft systems (UASs), known as counter UAS (cUAS), is pivotal in restricted air spaces. The application of deep learning (DL) classifiers on electro-optical (EO) sensor data is promising for cUAS, but it introduces three key challenges. Specifically, DL-based cUAS produces point estimates at test time with no associated measure of uncertainty (softmax outputs produced by typical DL models are often overconfident predictions, resulting in unreliable measures of uncertainty), easily triggers false positive detections for birds and other aerial wildlife, and cannot accurately characterize out-of-distribution (OOD) input samples. In this work, we develop an epistemic uncertainty quantification (UQ) framework, which utilizes the advantages of DL while simultaneously producing uncertainty estimates on both in-distribution and OOD input samples. In this context, in-distribution samples refer to testing samples collected according to the same data generation process as the training data, and OOD samples refer to in-distribution samples that are intentionally perturbed in order to shift the distribution of the testing set away from the distribution of the training set. Our framework produces a distributive estimate of each prediction, which accurately expresses UQ, as opposed to a point estimate produced by standard DL. Through evaluation on a custom field-collected dataset consisting of images captured from EO sensors and in comparison to prior cUAS baselines, we show that our framework effectively expresses low and high uncertainty on in-distribution and OOD samples, respectively, while retaining accurate classification performance. [ABSTRACT FROM AUTHOR]

Many communications and sensing applications hinge on the detection of a signal in a noisy, interference-heavy environment. Signal processing theory yields techniques such as the generalized likelihood ratio test (GLRT) to perform detection when the received samples correspond to a linear observation model. Numerous practical applications exist, however, where the received signal has passed through a nonlinearity, causing significant performance degradation of the GLRT. In this work, we propose prepending the GLRT detector with a neural network classifier capable of identifying the particular nonlinear time samples in a received signal. We show that pre-processing received nonlinear signals using our trained classifier to eliminate excessively nonlinear samples (i) improves the detection performance of the GLRT on nonlinear signals and (ii) retains the theoretical guarantees provided by the GLRT on linear observation models for accurate signal detection. [ABSTRACT FROM AUTHOR]

Over the last several decades, structural health monitoring systems have grown into increasingly diverse applications. Structural health monitoring excels with large data sets that can capture the typical variability, novel events, and undesired degradation over time. As a result, the efficient storage and processing of these large, guided wave data sets have become a key feature for successful application of structural health monitoring. This article describes a series of investigations into the use of random projection theory to significantly reduce storage burdens and improve computational complexity while not significantly affecting common damage detection strategies. Random projections are used as a lossy compression scheme that approximately retains metrics of distance or similarity between data records. Random projection compression is evaluated using a large 1,440,000 measurement data set, which was collected over 5 months in an unprotected outdoor environment. Accurate damage detection, after the compression process, is achieved through correlation analysis and singular value decomposition. The results indicate consistent detection performance with over 95% of storage compression and more than a 477 times speed improvement in computational cost for singular value decomposition–based damage detection. [ABSTRACT FROM AUTHOR]

In this study we have developed a conjunction model, WGP, of discrete wavelet transform (DWT) and genetic programming (GP) for forecasting river floods when the only data available are the historical daily flows. DWT is used for denoising and smoothening the observed flow time series on which GP is implemented to get the next-day flood. The new model is compared with autoregressive (AR) and stand-alone GP models. All models are calibrated and tested on the Kosi River which is one of the most devastating rivers of the world with high and spiky monsoon flows, modeling of which poses a great challenge. With different inputs, 12 models, four in each class of WGP, GP and AR, are devised. The best performing WGP model, WGP4, with four previous daily flow rates as input, forecasts the Kosi floods with an accuracy of 87.9%, root mean square error of 123.9 m3/s and Nash-Sutcliffe coefficient of 0.993, the best performance indices among all the developed models. The extreme floods are also better simulated by the WGP models than by AR and GP models. [ABSTRACT FROM AUTHOR]

The paper demonstrates the efficiency of Wavelet Regression (WR) in estimating floods in rivers when the only data available is historical flow series. Discrete Wavelet Transform (DWT) decomposes the flow series into constituent wavelet components, i.e., approximations and details. A modified flow series is then constructed after removing the most fluctuating components and recombining other wavelet components. The modified flow series forms the input basis for WR implementation. Autoregressive (AR) models, developed for the comparison purpose, were implemented on the original flow series. A case study of developed models was made using monsoon flood data of the Kosi River at Birpur gauge site in the Bihar state of India. Based on various performance indices, it can be concluded that WR models forecast floods with greater accuracy than AR models. [ABSTRACT FROM AUTHOR]

The sedimentation behaviour of Getalsud Reservoir in the Jharkhand State of India was studied, and its useful life and performance forecasted. The quantity of sediment and its vertical distribution over different reservoir elevations have been estimated using standard methods. The study results indicate the provisions made for sediment deposit in the reservoir are sufficient, and the useful life of the reservoir will be longer than its design life of 100 years. The trap-efficiency of the reservoir is not anticipated to vary much until the year 2250, after which it is expected to rapidly decrease. In the first 100 years of impoundment, the reservoir loses 60% of its dead storage, 13% of its live storage and 24% of its gross storage. At its present sedimentation rate, Getalsud Reservoir may become fully obsolete by the year 2500. [ABSTRACT FROM AUTHOR]

This paper is an attempt to predict one-day ahead flood levels of a river using Neural Networks (NNs), when only available records are flood levels of previous days. The Basua of Kosi river, India, is selected as the focus site. Several NNs are constructed and floods are forecasted. The results found that the feed forward back propagation network with input layer consisting of past two days' flood levels and the hidden layer with four neurons to be the best performing NN model for forecasting the current flood level. The same datasets were utilized for prediction of flood levels from the five developed Autoregression models (ARs). The predicted results of flood levels by NNs are very satisfactory and acceptable as compared to ARs and seems to be a good alternative for flood forecasting. [ABSTRACT FROM AUTHOR]

In this paper, a simulation linked dynamic programming model is presented for determination of optimal sizes of reservoirs in a multiple-reservoir system. These reservoirs satisfy the individual as well as the combined water demands from the system at a minimum cost. Simulation is used for the determination of reservoirs' storage-yield-cost functions, thereafter, dynamic programming is applied on these functions to get the optimal sizes of the reservoirs. A practical application of the model is made to an existing multiple-reservoir system, namely the Damodar Valley Reservoir System of Eastern India. There is an indication of better storage allocation to reservoirs by this model compared to the design values, resulting in cost saving of Rs. 160 cr. [ABSTRACT FROM AUTHOR]

A new expression for the prediction of longitudinal dispersion coefficient in natural rivers, using genetic algorithms, is proposed. The expression uses hydraulic and geometric characteristics of rivers, which are readily available. For performance evaluation, using published field data, results of coefficient prediction by the new expression and by the other reported expressions are compared. According to various performance indices, it is concluded that the new formula predicts the longitudinal dispersion coefficient more accurately. Sensitive analysis performed on input parameters indicates the ratio of the cross-sectional mean velocity to the bottom shear velocity to be the most influencing parameter for accurate prediction of the longitudinal dispersion coefficient. [ABSTRACT FROM AUTHOR]