Spatial Extrapolation Research Articles

Artificial neural networks applied to landslide susceptibility: The effect of sampling areas on model capacity for generalization and extrapolation

Artificial neural networks (ANNs) have been used to identify areas susceptible to landslides and constitute one of the most widely used methods for this purpose. Several factors can interfere in the performance of the models and their resulting maps (especially sampling). This research evaluated the influence of sampling areas on landslide susceptibility modelling and the capacity for generalization and spatial extrapolation of data. Based on an inventory of landslide scars, distributed in five areas of southern Brazil, non-occurrence samples were defined by means of different buffers (2–40 km) in relation to the landslides in order to test the effect of the spatial distribution of the non-occurrence samples on the modeling results. A total of 16 morphometric attributes of the terrain (extracted from a digital elevation model) were used as input variables of the model. Multilayered network training was carried out using a backpropagation algorithm and accuracy was calculated by means of the Area Under the Receiver Operating Characteristic Curve (AUROC). Model accuracy was between 0.739 and 0.931. This variation was explained mainly by the buffer used. The susceptibility map resulting from the model of greater accuracy was obtained with a 40-km buffer in order to collect non-occurrence samples. The great distance between the occurrence and non-occurrence samples facilitates the modelling, since it increases the morphometric differences between the sampling groups. When we used samples from only one of the sample areas, the spatial extrapolation of the susceptibility map to the other areas showed high performance. We conclude that the ANN model for landslides susceptibility mapping can be extrapolated spatially, considering the limits of the geomorphological unit or numerical domain of the data.

Deep learning approaches for improving prediction of daily stream temperature in data‐scarce, unmonitored, and dammed basins

AbstractBasin‐centric long short‐term memory (LSTM) network models have recently been shown to be an exceptionally powerful tool for stream temperature (Ts) temporal prediction (training in one period and predicting in another period at the same sites). However, spatial extrapolation is a well‐known challenge to modelling Ts and it is uncertain how an LSTM‐based daily Ts model will perform in unmonitored or dammed basins. Here we compiled a new benchmark dataset consisting of >400 basins across the contiguous United States in different data availability groups (DAG, meaning the daily sampling frequency) with and without major dams, and studied how to assemble suitable training datasets for predictions in basins with or without temperature monitoring. For prediction in unmonitored basins (PUB), LSTM produced a root‐mean‐square error (RMSE) of 1.129°C and an R2 of 0.983. While these metrics declined from LSTM's temporal prediction performance, they far surpassed traditional models' PUB values, and were competitive with traditional models' temporal prediction on calibrated sites. Even for unmonitored basins with major reservoirs, we obtained a median RMSE of 1.202°C and an R2 of 0.984. For temporal prediction, the most suitable training set was the matching DAG that the basin could be grouped into (for example, the 60% DAG was most suitable for a basin with 61% data availability). However, for PUB, a training dataset including all basins with data was consistently preferred. An input‐selection ensemble moderately mitigated attribute overfitting. Our results indicate there are influential latent processes not sufficiently described by the inputs (e.g., geology, wetland covers), but temporal fluctuations can still be predicted well, and LSTM appears to be a highly accurate Ts modelling tool even for spatial extrapolation.

A gradient-enhanced sequential nonparametric data assimilation framework for soil moisture flow

Soil water content (SWC) is a vital variable in the hydrological cycle, while simulation of it often relies on resolving the soil water flow equation. To cope with the unavailability or poor quality of physical models, various nonparametric data assimilation (DA) schemes have been established. However, there tends to be two significant common challenges in such methods: (1) the difficulty in capturing locally changing behaviors of time series, especially the peaks and troughs, and (2) poor statistical interpolation capability in time and space. These two challenges are no doubt attributed to the complete renunciation of physical constraints. Unlike previous physics-informed approaches that incorporated physical governing equations and engineering control into the loss functions, this study attempts to introduce additional physical constraints from data gradient into the model-free DA framework. As a follow-up study of Wang et al. (2021), a gradient-enhanced version of nonparametric DA schemes (i.e., GE-EnKFGP) is proposed. The temporal (daily) and spatial (vertical) gradients of the SWC are merged into the construction of the unsaturated flow dynamical models based on the Gaussian process (GP), while the Kalman update formulation is used to reconcile real-time observations. With the aid of a series of real-world cases, the performance of the GE-EnKFGP was compared with the original EnKFGP and its gradient-based version (GB-EnKFGP), where the temporal gradients of the SWC were used as the proxy for the SWC as the GP output. The results showed that the enhancement of the gradient information in the GE-EnKFGP led to a better estimation than the initial EnKFGP due to its more accurate identification of multiple local extrema. This should be attributed to the mass conservation constraint hidden within the temporal gradients and the implicit constraint of the driving force (or upper boundary) from the spatial gradients. Spatial gradients of the SWC outperformed temporal ones under various application scenarios. The GB-EnKFGP and GE-EnKFGP exhibited superior performances in retrieving surface SWC than that in the deeper layer. Hence, an enhancement scheme using only the temporal gradients of the surface layer was recommended. In the context of spatial extrapolation, the assistance of spatial gradients yielded an improved estimate of the deeper SWC quite robustly through GP training and assimilation of easy-to-access surface data. However, the implementation of the GB-EnKFGP and temporal gradient-enhanced EnKFGP [i.e., GE-EnKFGP (t)] run the risk of triggering a performance collapse due to the delayed response of SWC profiles to rainfall events.

Predicting subsurface sonar observations with satellite-derived ocean surface data in the California Current Ecosystem.

Vessel-based sonar systems that focus on the water column provide valuable information on the distribution of underwater marine organisms, but such data are expensive to collect and limited in their spatiotemporal coverage. Satellite data, however, are widely available across large regions and provide information on surface ocean conditions. If satellite data can be linked to subsurface sonar measurements, it may be possible to predict marine life over broader spatial regions with higher frequency using satellite observations. Here, we use random forest models to evaluate the potential for predicting a sonar-derived proxy for subsurface biomass as a function of satellite imagery in the California Current Ecosystem. We find that satellite data may be useful for prediction under some circumstances, but across a range of sonar frequencies and depths, overall model performance was low. Performance in spatial interpolation tasks exceeded performance in spatial and temporal extrapolation, suggesting that this approach is not yet reliable for forecasting or spatial extrapolation. We conclude with some potential limitations and extensions of this work.

Spatial and Temporal Patterns of Ross River Virus in Queensland, 2001-2020.

Ross River virus (RRV), the most common human arbovirus infection in Australia, causes significant morbidity and substantial medical costs. About half of Australian cases occur in Queensland. We describe the spatial and temporal patterns of RRV disease in Queensland over the past two decades. RRV notifications, human population data, and weather data from 2001 to 2020 were analysed by the Statistical Area Level 2 (SA2) area. Spatial interpolation or linear extrapolation were used for missing weather values and the estimated population in 2020, respectively. Notifications and incidence rates were analysed through space and time. During the study period, there were 43,699 notifications in Queensland. The highest annual number of notifications was recorded in 2015 (6182), followed by 2020 (3160). The average annual incidence rate was 5 per 10,000 people and the peak period for RRV notifications was March to May. Generally, SA2 areas in northern Queensland had higher numbers of notifications and higher incidence rates than SA2 areas in southern Queensland. The SA2 areas with high incidence rates were in east coastal areas and western Queensland. The timely prediction may aid disease prevention and routine vector control programs, and RRV management plans are important for these areas.

Digital mapping of soil classes using spatial extrapolation with imbalanced data

Digital mapping of soil classes using the extrapolation approach is timesaving, economically cheap, and helps collecting soil data from areas with difficult sampling. However, it has not been explored widely enough for digital mapping of soil classes. This study seeks to evaluate and compare several machine learning and regression algorithms for the extrapolation of soil sub-groups. Also the issue of imbalanced number of observations was addressed and oversampling technique was applied on the minority soil class to improve the models performance. The study area is located in central north Iran with 84 and 72 soil profiles sampled in the donor and recipient areas, respectively. A set of various environmental covariates including remotely sensed data, digital elevation model derivatives and geomorphology map were used as explanatory variables for predicting soil classes. Results showed that among eleven investigated models, C5.0 decision tree (DT), random forest (RF) and multinomial logistic regression (MNL) had the highest overall accuracy of 46%, 42% and 38%, respectively, for the extrapolation of soil classes. Also the Kappa statistic values for these models were 0.30, 0.24 and 0.22, respectively. Oversampling of the minority soil class led to an increase in overall accuracy for some of the models with the highest ones being DT = 53% and RF = 50%. Also, the Kappa value for DT and RF models increased to 0.39 and 0.35, respectively. In addition, oversampling of the minority soil class led to the prevention of losing this class in the final map.

Comparison of velocity models computed using different velocity variants and algorithms on a 2D seismic profile for poststack time migration

This article presents a construction method of the velocity field for poststack time migration for 2D seismic calculated on the basis of interval velocities in boreholes and structural interpretation, as well as the results of poststack time migration based on this solution. Three velocity field models have been developed. The models used differ in the way of spatial interpolation and extrapolation in the adopted calculation grid in the depth domain, which was created on the basis of a structural interpretation of 2D seismic profiles. Three methods of interpolation and extrapolation were used: Gaussian distribution, kriging and moving average. The spatial distribution of the interval velocities in the boreholes was made using the Petrel software by Schlumberger. The interval velocities along the analyzed seismic profile were extracted from the computed spatial interval velocity models, and after conversion from the depth to the time domain, they were used for the poststack time migration. For comparison, poststack time migration was calculated for the same seismic profile based on the stacking velocities obtained in the seismic processing data as a result of velocity analyzes. The velocity field calculated on the basis of interval velocities and structural interpretation was used for the poststack time migration procedure performed with the Implicit FD Time Migration algorithm (finite difference), while the stacking velocities were used for the poststack time migration procedure performed with the Stolt and Kirchhoff algorithms in accordance with the technical conditions of correct operation of these algorithms. The selected percentage ranges of 60%, 100%, and 140% have been used for all velocity fields. Application of the element of directional velocity variation resulting from the spatial distribution of interval velocities in the boreholes to the velocity field for the poststack time migration allowed to obtain a better seismic image in relation to the one obtained as a result of applying the stacking velocities. The most reliable seismic image after poststack time migration was obtained for the velocity field calculated on the basis of the interval velocities with Gaussian distribution, using the finite difference algorithm with 60 percent value of the velocity field.

Spatial extrapolation of topographic models for mapping soil organic carbon using local samples

Accurate prediction of soil organic carbon (SOC) at fine scales is frequently limited by data availability but fine-scale maps of soil properties are often of great importance for local agricultural land management and watershed processes assessment. Terrain attributes derived from elevation have been commonly used predictor variables in SOC mapping. The aim of this study is to map SOC density (0–30 cm) using hybrid methods that combine regional topographic models and a residual kriging procedure using local samples from an area of interest and evaluate the effects of different sampling designs for residual kriging on final map quality. The regional topographic model was developed based on light detection and ranging (lidar)-derived topographic covariates using random forest to map SOC density on a cropland site intensively sampled for SOC (n = 224). To test ability to generate a local calibration of this regional model, a second intensively sampled cropland site (n = 225) was used to generate a reference SOC density map where residual kriging was implemented involving different numbers of local samples (n = 10–255) and different sample selection schemes: stratified random sampling (SRS) and spatial coverage sampling with close-pairs inclusions (SCS+). To generate a population of local sample sets, 500 repeats were conducted for each scheme from the reference map. By analyzing the distribution of relative mean error (RME) of the overall map for the second site with repeated selection of local samples, we found a smaller variation in RME for SCS+ than that for SRS for a given sample size. However, the E(RME) estimates for SCS+ slightly varied with number of local samples used while the E(RME) for SRS was stabilized at zero over that range in local sampling intensity. There was negligible difference between the distributions of RME for the two sampling schemes at individual points, except a finding of somewhat lower RME for SCS+ based point estimates. These results indicate that SCS+ can generate smaller variation in overall map quality while SRS can provide unbiased mean estimate for an area of interest. This study demonstrates the utility of extrapolating a regional topographic model by use of a limited number of new samples for local calibration and provides insight for practitioners who are interested in mapping soil carbon distribution in sparsely sampled areas of interest.

Temperature sensitivity of plant litter decomposition rate in China's forests

AbstractThe decomposition of litter from forest tree species is a fundamental process in the carbon (C) cycle of terrestrial ecosystems and is closely related to ongoing climate change. However, the spatial distribution of the forest litter decomposition rate (k) and its potential response to changing air temperature (temperature sensitivity) remain poorly understood. Here, we estimated the spatial pattern of forest plantkvalues in China by performing a random forest model based on 433 standardizedkvalues from 124 published studies. Nine potential predictors, including climate‐related factors, vegetation characteristics, and soil physical and chemical factors, were considered in the model. The results of spatial extrapolation indicated that the averagekvalue for China's forests was 0.53, and the mean annual temperature (MAT) was the most important factor. We also mapped the spatial pattern of the temperature sensitivity of thekvalue (Q10) by using a moving window method. We found that theQ10values had considerable variation (from 0.05 to 11.68, 95% confidence interval, CI) across forest types and regions. TheQ10values were lower in the warmer regions (primarily in the sub‐tropical evergreen forests) and higher in the semi‐humid regions (primarily in the temperate deciduous forests and borealLarixforests) than those in the other regions of China. However, theQ10values of the broadleaved forests were higher than those of coniferous forests. These results suggest that the temperature sensitivity of the litter decomposition rate will decline under the ongoing global warming. Changing patterns of precipitation will also affect not only the forest litter decomposition rate but also its temperature sensitivity.

Assessing Marginal Shallow-Water Bathymetric Information Content of Lidar Sounding Attribute Data and Derived Seafloor Geomorphometry

Shallow-water depth estimates from airborne lidar data might be improved by using sounding attribute data (SAD) and ocean geomorphometry derived from lidar soundings. Moreover, an accurate derivation of geomorphometry would be beneficial to other applications. The SAD examined here included routinely collected variables such as sounding intensity and fore/aft scan direction. Ocean-floor geomorphometry was described by slope, orientation, and pulse orthogonality that were derived from the depth estimates of bathymetry soundings using spatial extrapolation and interpolation. Four data case studies (CSs) located near Key West, Florida (United States) were the testbed for this study. To identify bathymetry soundings in lidar point clouds, extreme gradient boosting (XGB) models were fitted for all seven possible combinations of three variable suites—SAD, derived geomorphometry, and sounding depth. R2 values for the best models were between 0.6 and 0.99, and global accuracy values were between 85% and 95%. Lidar depth alone had the strongest relationship to bathymetry for all but the shallowest CS, but the SAD provided demonstrable model improvements for all CSs. The derived geomorphometry variables contained little bathymetric information. Whereas the SAD showed promise for improving the extraction of bathymetry from lidar point clouds, the derived geomorphometry variables do not appear to describe geomorphometry well.

Spatial extrapolation of early room impulse responses in local area using sparse equivalent sources and image source method

The room impulse response (RIR) is important in most acoustic applications, such as the design of concert halls and sound field control, because it characterizes the sound propagation. The measurement of RIRs at multiple points is challenging, as it requires a huge microphone array or repeating the experiment by microphone replacement. Several RIR interpolation and extrapolation methods of RIRs have been developed for obtaining RIRs from multiple measurement points efficiently. Extrapolation methods offer more efficient RIR measurement compared to interpolation. However, previous studies focused on extrapolation at frequencies below 1 kHz, and the extrapolation at higher frequencies was difficult. In this study, we propose an extrapolation method for RIRs of direct sound and early reflections in a local area using a small number of measurement points. The proposed method represents the RIRs around the microphones using superpositions of sparse equivalent sources located around the loudspeaker and image sources. We conducted a measurement experiment in an anechoic chamber to estimate RIRs around microphones using sound-reflecting boards. From the experimental results with 2.5 dimensional conditions, the proposed method achieved about above 10 dB of signal to noise ratio (SNR) near the microphone array from 0.5–8.5 kHz. For the extrapolation accuracy over the entire evaluation area (0.6 × 0.54 m2), the proposed method improved the SNR by about 5–6 dB compared to results using the plane wave decomposition.

SBAS-Aided GPS Positioning with an Extended Ionosphere Map at the Boundaries of WAAS Service Area

Space-based augmentation system (SBAS) provides correction information for improving the global navigation satellite system (GNSS) positioning accuracy in real-time, which includes satellite orbit/clock and ionospheric delay corrections. At SBAS service area boundaries, the correction is not fully available to GNSS users and only a partial correction is available, mostly satellite orbit/clock information. By using the geospatial correlation property of the ionosphere delay information, the ionosphere correction coverage can be extended by a spatial extrapolation algorithm. This paper proposes extending SBAS ionosphere correction coverage by using a biharmonic spline extrapolation algorithm. The wide area augmentation system (WAAS) ionosphere map is extended and its ionospheric delay error is compared with the GPS Klobuchar model. The mean ionosphere error reduction at low latitude is 52.3%. The positioning accuracy of the extended ionosphere correction method is compared with the accuracy of the conventional SBAS positioning method when only a partial set of SBAS corrections are available. The mean positioning error reduction is 44.8%, and the positioning accuracy improvement is significant at low latitude.

Crop-specific technology extrapolation domains for Brazil

ABSTRACT Climate and soil databases can be used for delineating homogenous environmental zones over large areas of agricultural land to support agricultural research planning and spatial extrapolation of the results of field experiments. In other countries, such framework has proven useful to improve the efficiency and impact of investments in agricultural research. Furthermore, this approach can also contribute to identify key locations with the greatest potential for spatial extrapolation in terms of crop production area with similar climate and soils. In this paper, soil and weather databases covering the main agricultural land in Brazil were developed, and crop-specific technology extrapolation domains (TEDs) were identified for maize, soybean, and sugarcane crops. These databases were useful and viable option for the generation of homogeneous zones maps for different crops and regions. Given the spatial scale of the database applied, it was possible to represent a large area in few TEDs, without losing information regarding climate and soil variability, facilitating the spatial extrapolation of technologies throughout Brazil. Few TEDs are needed to represent a large part of the area of sugarcane (32 TEDs to cover 51% of the area), maize (21 TEDs for 51% of the area), and soybean (17 TEDs for 50% of the total area). Such information may help several areas of the agricultural sector, such as supporting new research that works on areas with large scales, or areas not explored yet.

ИНДЕКС ПОЛЕВОГО ОБСЛЕДОВАНИЯ ТЕРРИТОРИИ ПРИ СРЕДНЕМАСШТАБНОМ КАРТОГРАФИРОВАНИИ

The quality of thematic mapping mostly depends on the representativeness of the field survey of the territory. In large-scale mapping geostatistical techniques can be used to design survey spatial sampling. In medium-scale mapping, the sampling scheme is traditionally based on experience, common views about the thematic diversity of the territory and available cartographic data. In the digital map-ping of thematic variables secondary data is used. There are mainly remote sensing data and derivatives of a digital elevation model. Spatial extrapolation of thematic information of points is somehow connected with the use of distance measures between studied and unstudied points in the space of secondary variables. This measure is proposed to be used as a numerical measure of the survey quality of the entire mapping space. This article provides numerical experiments to illustrate the behavior of this index, as well as proposes possible ways of using the index to control fieldwork in order to improve the spatial model of thematic variables.

Estimating Time Since the Last Stand-Replacing Disturbance (TSD) from Spaceborne Simulated GEDI Data: A Feasibility Study

Stand-level maps of past forest disturbances (expressed as time since disturbance, TSD) are needed to model forest ecosystem processes, but the conventional approaches based on remotely sensed satellite data can only extend as far back as the first available satellite observations. Stand-level analysis of airborne LiDAR data has been demonstrated to accurately estimate long-term TSD (~100 years), but large-scale coverage of airborne LiDAR remains costly. NASA’s spaceborne LiDAR Global Ecosystem Dynamics Investigation (GEDI) instrument, launched in December 2018, is providing billions of measurements of tropical and temperate forest canopies around the globe. GEDI is a spatial sampling instrument and, as such, does not provide wall-to-wall data. GEDI’s lasers illuminate ground footprints, which are separated by ~600 m across-track and ~60 m along-track, so new approaches are needed to generate wall-to-wall maps from the discrete measurements. In this paper, we studied the feasibility of a data fusion approach between GEDI and Landsat for wall-to-wall mapping of TSD. We tested the methodology on a ~52,500-ha area located in central Idaho (USA), where an extensive record of stand-replacing disturbances is available, starting in 1870. GEDI data were simulated over the nominal two-year planned mission lifetime from airborne LiDAR data and used for TSD estimation using a random forest (RF) classifier. Image segmentation was performed on Landsat-8 data, obtaining image-objects representing forest stands needed for the spatial extrapolation of estimated TSD from the discrete GEDI locations. We quantified the influence of (1) the forest stand map delineation, (2) the sample size of the training dataset, and (3) the number of GEDI footprints per stand on the accuracy of estimated TSD. The results show that GEDI-Landsat data fusion would allow for TSD estimation in stands covering ~95% of the study area, having the potential to reconstruct the long-term disturbance history of temperate even-aged forests with accuracy (median root mean square deviation = 22.14 years, median BIAS = 1.70 years, 60.13% of stands classified within 10 years of the reference disturbance date) comparable to the results obtained in the same study area with airborne LiDAR.

Balancing transferability and complexity of species distribution models for rare species conservation

AbstractAimSpecies distribution models (SDMs) are valuable for rare species conservation and are commonly used to extrapolate predictions of habitat suitability geographically to regions where species occurrence is unknown (i.e., transferability). Spatially structured cross‐validation can be used to infer transferability, yet, few studies have evaluated how delineation of cross‐validation folds affects model complexity and predictions. We developed SDMs using multiple cross‐validation approaches to understand the implications for predicting habitat suitability for northern Idaho ground squirrels, a rare, federally threatened species that has been extensively surveyed in regions where known populations occur, resulting in >8000 presence locations.LocationIdaho, USA.MethodsWe delineated cross‐validation folds by mimicking the manner in which predictions would be geographically extrapolated or by using existing dispersal barriers. We varied the distance between, number, and directionality of folds. We conducted a grid search on statistical regularization parameters to optimize model complexity, covering a range of values exceeding that typically implemented. For each cross‐validation approach, we selected optimal regularization and model complexity based on out‐of‐sample predictive ability.ResultsDelineation of cross‐validation folds substantially affected resulting model complexity and extrapolated predictions. All cross‐validation approaches resulted in models with apparently high out‐of‐sample predictive ability, yet optimal model complexity varied substantially among the approaches. Regularization demonstrated a noisy relationship between model complexity and prediction, where local optima in predictive performance were common at small values.Main conclusionSubtle modelling decisions can have large consequences for predictions of habitat suitability and transferability of SDMs. When transferability is the goal, cross‐validation approaches should be considered carefully and mimic the manner in which spatial extrapolation will occur, else overly complex models with inflated assessments of predictive accuracy may result. Further, spatially structured cross‐validation may not guard against over‐parameterization, and assessing a broader range of regularization parameters may be necessary to optimize model complexity for transferability.

Comparison of Different Missing-Imputation Methods for MAIAC (Multiangle Implementation of Atmospheric Correction) AOD in Estimating Daily PM2.5 Levels

The immense problem of missing satellite aerosol retrievals (Aerosol Optical Depth, (AOD)) detrimentally affects the prediction ability of ground-level PM2.5 concentrations and may lead to unavoidable biases. An appropriate missing-imputation method has not been well developed to date. This study developed a two-stage approach (AOD-imputation stage and PM2.5-prediction stage) to predict short-term PM2.5 exposure in mainland China from 2013–2018. At the AOD-imputation stage, geostatistical methods and machine learning (ML) algorithms were examined to interpolate 1 km satellite aerosol retrievals. At the PM2.5-prediction stage, the daily levels of PM2.5 were predicted at a resolution of 1 km, based on interpolated AOD and meteorological data. The statistical performances of the different interpolation methods were comprehensively compared at each stage. The original coverage of retrieved AOD was 15.46% on average. For the AOD-imputation stage, ML methods produced a higher coverage (98.64%) of AOD than geostatistical methods (21.43–87.31%). Among ML algorithms, random forest (RF) or extreme gradient boosted (XG-interpolated) AOD produced better interpolated quality (CV R2 = 0.89 and 0.85) than other algorithms (0.49–0.78), but XGBoost required only 15% of the computing time of RF. For the PM2.5 predicted stage, neither RF-AOD nor XG-AOD could guarantee higher accuracy in PM2.5 estimations (CV R2 = 0.88 (RF or XG-AOD) compared to 0.85 (original)), or more stable spatial and temporal extrapolation (spatial, (temporal) CV R2 = 0.83 (0.83), 0.82 (0.82), and 0.65 (0.61) for RF, XG, and original). For the AOD-imputation stage, the missing-filled efficiency depended more on external information, while the missing-filled accuracy relied more on model structure. For the PM2.5 predicted stage, efficient AOD interpolation (or the ability to eliminate the missing data) was a precondition for the stable spatial and temporal extrapolation, while the quality of interpolated AOD showed less significant improvements. It was found that XG-AOD is a better choice to estimate daily PM2.5 exposure in health assessments.

Soft Sensor Modeling of Blast Furnace Wall Temperature Based on Temporal–Spatial Dimensional Finite-Element Extrapolation

The blast furnace (BF) wall temperature field plays a key role in determining whether the operation of the BF is efficient. However, it cannot be measured directly because the BF is an enclosed complex reactor with an extremely harsh environment. To cope with this problem, a soft sensor modeling strategy of the BF wall temperature field based on a novel temporal–spatial dimensional finite-element extrapolation algorithm (TS-FEEA) is designed. A 3-D transient heat conduction model of the BF wall is first established, which considers the layered physical structure of the BF. The large volume of the BF wall causes great calculation burden when solving the 3-D transient heat conduction model of the BF wall, to solve this problem, a TS-FEEA is proposed that realizes a high-speed and accurate model numerical solution. To further improve the 3-D model precision, the measured thermocouple temperatures in the cooling stave are served as the judgment criteria, and a correction strategy for inner wall temperature is finally designed to correct the assumed inner wall temperature through multiple iterations. Simulation results demonstrate the validity of the soft sensor modeling strategy and its applicability in industrial plants. The research achievements provide valuable reference information for the BF operators to judge the furnace conditions and adjust the operation parameters.

Peak Ground Velocity Spatial Variability Revealed by Dense Seismic Array in Southern California

AbstractUnderstanding and modeling variability of ground motion is essential for building accurate and precise ground motion prediction equations, which can net site‐specific characterization and reduced hazard levels. Here, we explore the spatial variability in peak ground velocity (PGV) at Sage Brush Flats along the San Jacinto Fault in Southern California. We use data from a dense array (0.6 × 0.6 km2, 1,108 geophones, station spacings 10–30 m) deployed in 2014 for ~1 month. These data offer an opportunity to study small‐scale variability in this region. We examine 38 earthquakes (2 ≤ ML ≤ 4.2) within 200 km of the array. Fault strands and a small basin impact the ground motions, producing PGV variations up to 22% of the mean and a 40% reduction in P and S wave near‐surface velocities. We find along‐fault rupture directivity, source, and path effects can increase PGVs by 167%. Surface PGV measurements exceed the colocated borehole station (depth at 148 m) PGV by factors of 3–10, confirming the impact on PGV from near‐surface fault structures, basins, topography, and amplifications from soft sediments. Consistently, we find high PGVs within the basin structure. A pair of colocated GaML2.6 events produce repeatable PGV values with similar spatial patterns. The average corner frequencies of these two events are 11–16 Hz, and viable measurements of stress drop can differ by 6.45 MPa. Within this small array, the PGV values are variable implying spatial extrapolation of PGV to regions of known faults and basins, even across a small area, should be done with caution.

Assessing the reliability of predicted plant trait distributions at the global scale.

AimPredictions of plant traits over space and time are increasingly used to improve our understanding of plant community responses to global environmental change. A necessary step forward is to assess the reliability of global trait predictions. In this study, we predict community mean plant traits at the global scale and present a systematic evaluation of their reliability in terms of the accuracy of the models, ecological realism and various sources of uncertainty.LocationGlobal.Time periodPresent.Major taxa studiedVascular plants.MethodsWe predicted global distributions of community mean specific leaf area, leaf nitrogen concentration, plant height and wood density with an ensemble modelling approach based on georeferenced, locally measured trait data representative of the plant community. We assessed the predictive performance of the models, the plausibility of predicted trait combinations, the influence of data quality, and the uncertainty across geographical space attributed to spatial extrapolation and diverging model predictions.ResultsEnsemble predictions of community mean plant height, specific leaf area and wood density resulted in ecologically plausible trait–environment relationships and trait–trait combinations. Leaf nitrogen concentration, however, could not be predicted reliably. The ensemble approach was better at predicting community trait means than any of the individual modelling techniques, which varied greatly in predictive performance and led to divergent predictions, mostly in African deserts and the Arctic, where predictions were also extrapolated. High data quality (i.e., including intraspecific variability and a representative species sample) increased model performance by 28%.Main conclusionsPlant community traits can be predicted reliably at the global scale when using an ensemble approach and high‐quality data for traits that mostly respond to large‐scale environmental factors. We recommend applying ensemble forecasting to account for model uncertainty, using representative trait data, and more routinely assessing the reliability of trait predictions.