Open land cover from OpenStreetMap and remote sensing

OpenStreetMap (OSM) tags were used to produce a global Open Land Cover (OLC) product with fractional data gaps available at osmlanduse.org. Data gaps in the global OLC map were filled for a case study in Heidelberg, Germany using free remote sensing data, which resulted in a land cover (LC) prototype with complete coverage in this area. Sixty tags in the OSM were used to allocate a Corine Land Cover (CLC) level 2 land use classification to 91.8% of the study area, and the remaining gaps were filled with remote sensing data. For this case study, complete are coverage OLC overall accuracy was estimated 87%, which performed better than the CLC product (81% overall accuracy) of 2012. Spatial thematic overlap for the two products was 84%. OLC was in large parts found to be more detailed than CLC, particularly when LC patterns were heterogeneous, and outperformed CLC in the classification of 12 of the 14 classes. Our OLC product represented data created in different periods; 53% of the area was 2011–2016, and 46% of the area was representative of 2016–2017.     

基于随机森林的PM2.5实时预报系统

针对我国当前重污染天气PM2.5浓度的实时预报问题,该文提出了一种基于随机森林算法的PM2.5浓度实时预报方法,并利用此方法对北京市地面空气质量监测数据和气象数据进行分析,建立了基于随机森林算法的PM2.5浓度实时预报模型。实验证明,该模型能够对72h内PM2.5浓度进行较高精度的实时预报,通过使用Spark分布式计算框架,能够有效降低算法耗时,文章基于此模型与Spark分布式计算框架建立了PM2.5实时预报系统。     

非均质空间动态随机扩散的一般方程及其应用

通过非均质空间条件下革新随机游动模式的理论推导，结合实证研究成果，建立了非均质空间动态随机扩散方程。同时，将此方程应用于城市基准地价评估，予以实例验证。     

Three-dimensional slope reliability and risk assessment using auxiliary random finite element method

This paper aims to propose an auxiliary random finite element method (ARFEM) for efficient three-dimensional (3-D) slope reliability analysis and risk assessment considering spatial variability of soil properties. The ARFEM mainly consists of two steps: (1) preliminary analysis using a relatively coarse finite-element model and Subset Simulation, and (2) target analysis using a detailed finite-element model and response conditioning method. The 3-D spatial variability of soil properties is explicitly modeled using the expansion optimal linear estimation approach. A 3-D soil slope example is presented to demonstrate the validity of ARFEM. Finally, a sensitivity study is carried out to explore the effect of horizontal spatial variability. The results indicate that the proposed ARFEM not only provides reasonably accurate estimates of slope failure probability and risk, but also significantly reduces the computational effort at small probability levels. 3-D slope probabilistic analysis (including both 3-D slope stability analysis and 3-D spatial variability modeling) can reflect slope failure mechanism more realistically in terms of the shape, location and length of slip surface. Horizontal spatial variability can significantly influence the failure mode, reliability and risk of 3-D slopes, especially for long slopes with relatively strong horizontal spatial variability. These effects can be properly incorporated into 3-D slope reliability analysis and risk assessment using ARFEM.     

Assessing the social context of wildfire-affected areas. The case of mainland Portugal

Wildfires cause different impacts, depending on the conditions and resilience level of the exposed communities. Wildfire occurrence in mainland Portugal was assessed with regard to socioeconomic and demographic parameters, to identify the most distinctive conditions of fire-affected areas, without implying the existence of causal relationships. The latest population and agriculture census data were used to retrieve conditions at the civil parish level, regarding demographic patterns, social and labor conditions, physical structures and agricultural activities. To identify differences between parishes, two groups were created with the communities that showed the highest and lowest 20% of wildfire incidence between 2007 and 2014, separately for density of fire events and for burned area. A stepwise approach based on classification trees and random Forest methods was applied to identify the best discriminant variables between the groups. First, irrelevant variables were removed by an interactive process based on misclassification rates. The second step used random Forest analysis to the remaining variables to evaluate their importance in distinguishing the groups. In the final step, cluster analysis was applied to test the correspondence between the clusters created with the selected variables and the initial groups. Results showed that parishes with higher fire density have higher population density, higher proportion of young and educated people, larger families and more overcrowded buildings. On the contrary, parishes with larger burned area are less populated, less attractive to foreigners, have a higher proportion of elderly people, more degraded housing conditions and agricultural activities, visible in the density of sheep and goat and pastures, are still relevant. The cluster analysis demonstrated a better performance of the model for wildfire density, revealing a strong association with socioeconomic dynamics with an agreement above 0.85, much higher than for burned areas which is 0.29. Overall, the spatial distribution of wildfire impacts is framed by societal settings and particular conditions must be further understood to improve the coping capacity of affected communities.     

A systematic comparison of different object-based classification techniques using high spatial resolution imagery in agricultural environments

Geographic Object-Based Image Analysis (GEOBIA) is becoming more prevalent in remote sensing classification, especially for high-resolution imagery. Many supervised classification approaches are applied to objects rather than pixels, and several studies have been conducted to evaluate the performance of such supervised classification techniques in GEOBIA. However, these studies did not systematically investigate all relevant factors affecting the classification (segmentation scale, training set size, feature selection and mixed objects). In this study, statistical methods and visual inspection were used to compare these factors systematically in two agricultural case studies in China. The results indicate that Random Forest (RF) and Support Vector Machines (SVM) are highly suitable for GEOBIA classifications in agricultural areas and confirm the expected general tendency, namely that the overall accuracies decline with increasing segmentation scale. All other investigated methods except for RF and SVM are more prone to obtain a lower accuracy due to the broken objects at fine scales. In contrast to some previous studies, the RF classifiers yielded the best results and the k-nearest neighbor classifier were the worst results, in most cases. Likewise, the RF and Decision Tree classifiers are the most robust with or without feature selection. The results of training sample analyses indicated that the RF and adaboost. M1 possess a superior generalization capability, except when dealing with small training sample sizes. Furthermore, the classification accuracies were directly related to the homogeneity/heterogeneity of the segmented objects for all classifiers. Finally, it was suggested that RF should be considered in most cases for agricultural mapping.     

Timber production assessment of a plantation forest: An integrated framework with field-based inventory,multi-source remote sensing data and forest management history

Timber production is the purpose for managing plantation forests, and its spatial and quantitative information is critical for advising management strategies. Previous studies have focused on growing stock volume (GSV), which represents the current potential of timber production, yet few studies have investigated historical process-harvested timber. This resulted in a gap in a synthetical ecosystem service assessment of timber production. In this paper, we established a Management Process–based Timber production (MPT) framework to integrate the current GSV and the harvested timber derived from historical logging regimes, trying to synthetically assess timber production for a historical period. In the MPT framework, age-class and current GSV determine the times of historical thinning and the corresponding harvested timber, by using a “space-for-time” substitution. The total timber production can be estimated by the historical harvested timber in each thinning and the current GSV. To test this MPT framework, an empirical study on a larch plantation (LP) with area of 43,946 ha was conducted in North China for a period from 1962 to 2010. Field-based inventory data was integrated with ALOS PALSAR (Advanced Land-Observing Satellite Phased Array L-band Synthetic Aperture Radar) and Landsat-8 OLI (Operational Land Imager) data for estimating the age-class and current GSV of LP. The random forest model with PALSAR backscatter intensity channels and OLI bands as input predictive variables yielded an accuracy of 67.9% with a Kappa coefficient of 0.59 for age-class classification. The regression model using PALSAR data produced a root mean square error (RMSE) of 36.5 m³ ha⁻¹. The total timber production of LP was estimated to be 7.27 × 10⁶ m³, with 4.87 × 10⁶ m³ in current GSV and 2.40 × 10⁶ m³ in harvested timber through historical thinning. The historical process-harvested timber accounts to 33.0% of the total timber production, which component has been neglected in the assessments for current status of plantation forests. Synthetically considering the RMSE for predictive GSV and misclassification of age-class, the error in timber production were supposed to range from −55.2 to 56.3 m³ ha⁻¹. The MPT framework can be used to assess timber production of other tree species at a larger spatial scale, providing crucial information for a better understanding of forest ecosystem service.     

Probability distribution of random wave–current forces

Based on the second-order solutions obtained for the three-dimensional weakly nonlinear random waves propagating over a steady uniform current in finite water depth, the joint statistical distribution of the velocity and acceleration of the fluid particle in the current direction is derived using the characteristic function expansion method. From the joint distribution and the Morison equation, the theoretical distributions of drag forces, inertia forces and total random forces caused by waves propagating over a steady uniform current are determined. The distribution of inertia forces is Gaussian as that derived using the linear wave model, whereas the distributions of drag forces and total random forces deviate slightly from those derived utilizing the linear wave model. The distributions presented can be determined by the wave number spectrum of ocean waves, current speed and the second order wave–wave and wave–current interactions. As an illustrative example, for fully developed deep ocean waves, the parameters appeared in the distributions near still water level are calculated for various wind speeds and current speeds by using Donelan–Pierson–Banner spectrum and the effects of the current and the nonlinearity of ocean waves on the distribution are studied.     

Scour below pipelines and around vertical piles in random waves

An approach by which the scour depth and scour width below a fixed pipeline and scour depth around a circular vertical pile in random waves can be derived is presented. Here, the scour depth formulas by Sumer and Fredsøe [ASCE J. Waterw., Port, Coastal Ocean Eng. 116 (1990) 307] for pipelines and Sumer et al. [ASCE J. Waterw., Port, Coastal Ocean Eng. 114 (1992) 599] for vertical piles as well as the scour width formula by Sumer and Fredsøe [The Mechanics of Scour in the Marine Environment, World Scientific, Singapore, 2002] for pipelines combined with describing the waves as a stationary Gaussian narrow-band random process are used to derive the cumulative distribution functions of the scour depths and width. Comparisons are made between the present approach and random wave scour data. Tentative approaches to related random wave scour cases are also suggested.     

Acoustic plane-wave reflection from a rough surface over a random fluid half-space

Plane-wave reflection from a rough surface overlying a fluid half-space, with a sound speed distribution subject to a small and random perturbation, is considered. A theory based upon a boundary perturbation method in conjunction with a formulation derived from Green's function for the coherent field in the random medium have been applied to a typical oceanic environment to study their effects on the plane-wave reflection. By considering the coherent field itself, the plane-wave reflection may be obtained straightforwardly through a procedure consistent with the formalisms currently employed in rough surface scattering. The results show that both the rough surface and medium inhomogenieties may reduce the plane-wave reflection, however, the characteristics of the curves representing their effects are different, enabling us to identify the dominant scattering mechanism. The results for the coherent reflection due to the individual scattering mechanism are compatible with those found in the existing literature.