Journal Paper Digests 2017 #3
Hybrid integration of Multilayer Perceptron Neural Networks and machine learning ensembles for landslide susceptibility assessment at Himalayan area (India) using GIS
Authors: Pham, BT; Bui, DT; Prakash, I; Dholakia, MB
Source: CATENA, 149 52-63; 1 FEB 2017
Abstract: The main objective of this study is to evaluate and compare the performance of landslide models using machine learning ensemble technique for landslide susceptibility assessment. This technique is a combination of ensemble methods (AdaBoost, Bagging, Dagging, MultiBoost, Rotation Forest, and Random SubSpace) and the base classifier of Multiple Perceptron Neural Networks (MLP Neural Nets). Ensemble techniques have been widely applied in other fields; however, their application is still rare in the assessment of landslide problems. Meanwhile, MLP Neural Nets, which is known as an artificial neural network, has been applied widely and efficiently in landslide problems. In the present study, landslide models of part Himalayan area (India) have been constructed and validated. For the evaluation and comparison of these models, receiver operating characteristic curve and Chi Square test methods have been applied. Overall, all landslide models performed well in landslide susuceptibility assessment but the performance of the MultiBoost model is the highest (AUC = 0.886), followed by Dagging model (AUC = 0.885), the Rotation Forest model (AUC = 0.882), the Bagging and Random SubSpace models (AUC = 0.881), and the AdaBoost model (AUC = 0.876), respectively. Moreover, machine learning ensemble models have improved significantly the performance of the base classifier of MLP Neural Nets (AUC = 0.874). Analysis of results indicates that landslide models using machine learning ensemble frameworks are promising methods which can be used as alternatives of individual base classifiers for landslide susceptibility assessment of other prone areas.
Ecosystem responses to land abandonment in Western Mediterranean Mountains
Authors: Romero-Diaz, A; Ruiz-Sinoga, JD; Robledano-Aymerich, F; Brevik, EC; Cerda, A
Source: CATENA, 149 824-835; SI 3 FEB 2017
Abstract: Agricultural expansion in the Mediterranean resulted in plant and soil degradation due to the intensive use, climate conditions, and rugged terrain. After abandonment, the recovery of vegetation contributed to improvement in soil quality from a hydrological, pedological and geomorphological point of view. This paper shows three examples of ecosystem evolution in abandoned fields in Valencia, Murcia and Andalucia and the application of different methodological approaches that resulted in similar findings. In Valencia, the main responses were the recovery of vegetation after land abandonment and an increase in organic matter and infiltration capacity of soils. In Murcia, with the exception of some terraced areas on marls, where erosion processes following abandonment were important, land abandonment resulted in vegetation recovery, improved soil properties, and reduced surface wash and soil losses. In Andalucia, research along climatological gradients showed the relationship between vegetation patterns and soil moisture and the control that climate exerts on hydrological and erosive behaviour. The experimental research conducted in three different regions in Western Mediterranean demonstrated that abandonment can result in recovery of the geo-ecosystem as vegetation and soil quality improvements were shown. The marls areas in Murcia were the exception with low soil quality and low vegetation cover, and as a consequence showed evidence of high erosion rates after abandonment.
Hydropedology and the Societal Challenge of Realizing the 2015 United Nations Sustainable Development Goals
Authors: Bouma, J
Source: VADOSE ZONE JOURNAL, 15 (12):NIL_57-NIL_63; DEC 2016
Abstract: The UN Sustainable Development Goals (SDGs) offer a major challenge for both society and the science community. Hydropedology, combining the expertise of soil physicists and pedologists, plays a key role in realizing goals focused on food, water, climate, and ecology, requiring interdisciplinary research. This update explores emerging trends and future work, focusing on examples of contributions by pedology to measuring and modeling in hydropedological studies. Many soil types create heterogeneous flow patterns that are difficult to characterize using current soil databases and physical flow models. The clear potential of hydropedology to produce better modeling results than those obtained from separate contributions by the two subdisciplines can, however, only be established by field validation of modeling results using different types of data and models. Overall soil input in interdisciplinary SDG-oriented research includes chemical and biological aspects that become more representative by considering hydropedological conditions.
Processes and Modeling of Initial Soil and Landscape Development: A Review
Authors: Maurer, T; Gerke, HH
Source: VADOSE ZONE JOURNAL, 15 (12):NIL_3-NIL_30; DEC 2016
Abstract: The development pathway of a landscape depends to a degree on the initial spatial distributions of mineral and organic components. The interaction between structures and ecohydrological processes during the critical initial development period is scarcely understood and often difficult to observe. While viable modeling approaches exist for most aspects of initial development (e.g., landscape evolution, vegetation succession), a deeper understanding of the prevailing feedback mechanisms requires a comprehensive, integrated modeling approach. We present a review of the current literature regarding the description of initial structures, the state-of-the-art of research on structure-forming processes and their interaction with existing and newly emerging structures, as well as the corresponding modeling efforts. The most relevant aspects are (i) sediment translocation processes and initial evolution of topography, (ii) surface crusting, (iii) vegetation succession, and (iv) the evolution of the soil pore space. Based on existing conceptions for integrated modeling of the coevolution of structures and ecohydrological behavior, we outline an integrative modeling framework that is based on a three-dimensional spatial structural model of initial sediment distributions, which can be used to: (i) analyze the spatiotemporal development dynamics depending on initial structures; and (ii) relate the simulated structural development to available observations of initial ecohydrological development. We discuss possible validation and generalization strategies of modeling results, and propose to define three-dimensional spatial functional catchment units (process domains) characterized by specific structural dynamics and the dominant ecohydrological processes.
From robustness to resilience: A network Price identity approach
Authors: Dragicevic, AZ
Source: ECOLOGICAL COMPLEXITY, 28 47-53; DEC 2016
Abstract: We model a closed-loop network of agents distributed among subnetworks and study the sustainability of network structures in presence of random perturbations. The model outcomes show that the stability of compartmentalized networks built on uniform operators depends on perturbations on between-subnetwork coupling, while the stability of networks built on mutation operators depends on their assimilation capacity. Through the study of eigenvalues of the Laplacian, we succeed in measuring the degree of network robustness and resilience. Our results also permit to situate the Price theorem, both in its standard and expanded forms, in the context of network evolutionary variational identity.
Fuzzy based energy efficient sensor network protocol for precision agriculture
Authors: Maurya, S; Jain, VK
Source: COMPUTERS AND ELECTRONICS IN AGRICULTURE, 130 20-37; NOV 15 2016
Abstract: Today’s, wireless sensor network have become a more emerging technology in precision agriculture. This paper proposes a novel approach based on sensor network technology to control the irrigation in agricultural field automatically. All sensor nodes deployed in the field, continuously sense soil temperature, soil moisture and air humidity of the agricultural field and transmit this information to base station only when the user defined periodic timer or sensed attributes values exceed desired threshold. In the proposed routing protocol, region-based static clustering approach is used to provide efficient coverage over entire agricultural area and threshold sensitive hybrid routing is used for transmitting sensed data to base station. The proposed protocol uses fuzzy logic technique to select the best cluster head among other sensor nodes in a particular round which minimizes the energy consumption of nodes in every data transmission period. The proposed energy-efficient protocol is compared with existing benchmark EEHC, DEEC, DDEEC and RBHR routing protocols. The analysis and experimental results show a significant decrement in data transmission rate due to user-defined transmission thresholds. The balanced use of fuzzy logic technique, static clustering and hybrid routing approach efficiently reduce energy consumption of sensor nodes in every data transmission round and prolongs the overall network lifetime by 173.16%, 149.22%, 99.49% and 4739% as compared to EEHC, DEEC, DDEEC and RBHR protocol respectively.
Mapping almond orchard canopy volume, flowers, fruit and yield using lidar and vision sensors
Authors: Underwood, JP; Hung, C; Whelan, B; Sukkarieh, S
Source: COMPUTERS AND ELECTRONICS IN AGRICULTURE, 130 83-96; NOV 15 2016
Abstract: This paper present a mobile terrestrial scanning system for almond orchards, that is able to efficiently map flower and fruit distributions and to estimate and predict yield for individual trees. A mobile robotic ground vehicle scans the orchard while logging data from on-board lidar and camera sensors. An automated software pipeline processes the data offline, to produce a 3D map of the orchard and to automatically detect each tree within that map, including correct associations for the same trees seen on prior occasions. Colour images are also associated to each tree, leading to a database of images and canopy models, at different times throughout the season and spanning multiple years. A canopy volume measure is derived from the 3D models, and classification is performed on the images to estimate flower and fruit density. These measures were compared to individual tree harvest weights to assess the relationship to yield. A block of approximately 580 trees was scanned at peak bloom, fruit-set and just before harvest for two subsequent years, with up to 50 trees individually harvested for comparison. Lidar canopy volume had the strongest linear relationship to yield with R-2 = 0.77 for 39 tree samples spanning two years. An additional experiment was performed using hand-held photography and image processing to measure fruit density, which exhibited similar performance (R-2 = 0.71). Flower density measurements were not strongly related to yield, however, the maps show clear differentiation between almond varieties and may be useful for other studies