Learning-Based Seismic Velocity Inversion with Synthetic and Field Data.

Building accurate acoustic subsurface velocity models is essential for successful industrial exploration projects. Traditional inversion methods from field-recorded seismograms struggle in regions with complex geology. While deep learning (DL) presents a promising alternative, its robustness using field data in these complicated regions has not been sufficiently explored. In this study, we present a thorough analysis of DL's capability to harness labeled seismograms, whether field-recorded or synthetically generated, for accurate velocity model recovery in a challenging region of the Gulf of Mexico. Our evaluation centers on the impact of training data selection and data augmentation techniques on the DL model's ability to recover velocity profiles. Models trained on field data produced superior results to data obtained using quantitative metrics like Mean Squared Error (MSE), Structural Similarity Index Measure (SSIM), and R2 (R-squared). They also yielded more geologically plausible predictions and sharper geophysical migration images. Conversely, models trained on synthetic data, while less precise, highlighted the potential utility of synthetic training data, especially when labeled field data are scarce. Our work shows that the efficacy of synthetic data-driven models largely depends on bridging the domain gap between training and test data through the use of advanced wave equation solvers and geologic priors. Our results underscore DL's potential to advance velocity model-building workflows in industrial settings using previously labeled field-recorded seismograms. They also highlight the indispensable role of earth scientists' domain expertise in curating synthetic data when field data are lacking.

Methodological considerations for studying neural oscillations.

Neural oscillations are ubiquitous across recording methodologies and species, broadly associated with cognitive tasks, and amenable to computational modelling that investigates neural circuit generating mechanisms and neural population dynamics. Because of this, neural oscillations offer an exciting potential opportunity for linking theory, physiology and mechanisms of cognition. However, despite their prevalence, there are many concerns-new and old-about how our analysis assumptions are violated by known properties of field potential data. For investigations of neural oscillations to be properly interpreted, and ultimately developed into mechanistic theories, it is necessary to carefully consider the underlying assumptions of the methods we employ. Here, we discuss seven methodological considerations for analysing neural oscillations. The considerations are to (1) verify the presence of oscillations, as they may be absent; (2) validate oscillation band definitions, to address variable peak frequencies; (3) account for concurrent non-oscillatory aperiodic activity, which might otherwise confound measures; measure and account for (4) temporal variability and (5) waveform shape of neural oscillations, which are often bursty and/or nonsinusoidal, potentially leading to spurious results; (6) separate spatially overlapping rhythms, which may interfere with each other; and (7) consider the required signal-to-noise ratio for obtaining reliable estimates. For each topic, we provide relevant examples, demonstrate potential errors of interpretation, and offer suggestions to address these issues. We primarily focus on univariate measures, such as power and phase estimates, though we discuss how these issues can propagate to multivariate measures. These considerations and recommendations offer a helpful guide for measuring and interpreting neural oscillations.

Dispersal in heterogeneous environments drives population dynamics and control of tsetse flies.

Spatio-temporally heterogeneous environments may lead to unexpected population dynamics. Knowledge is needed on local properties favouring population resilience at large scale. For pathogen vectors, such as tsetse flies transmitting human and animal African trypanosomosis, this is crucial to target management strategies. We developed a mechanistic spatio-temporal model of the age-structured population dynamics of tsetse flies, parametrized with field and laboratory data. It accounts for density- and temperature-dependence. The studied environment is heterogeneous, fragmented and dispersal is suitability-driven. We confirmed that temperature and adult mortality have a strong impact on tsetse populations. When homogeneously increasing adult mortality, control was less effective and induced faster population recovery in the coldest and temperature-stable locations, creating refuges. To optimally select locations to control, we assessed the potential impact of treating them and their contribution to the whole population. This heterogeneous control induced a similar population decrease, with more dispersed individuals. Control efficacy was no longer related to temperature. Dispersal was responsible for refuges at the interface between controlled and uncontrolled zones, where resurgence after control was very high. The early identification of refuges, which could jeopardize control efforts, is crucial. We recommend baseline data collection to characterize the ecosystem before implementing any measures.

Richardson's ground squirrel litter size-sex ratio trade-off reveals conditional adaptive sex allocation.

Trivers and Willard proposed that female mammals should adjust their investment in male versus female offspring relative to their ability to produce high-quality offspring. We tested whether litter size-sex ratio trade-offs predicted by Adaptive Sex Allocation (ASA) theory occur among Richardson's ground squirrel (Urocitellus richardsonii) dams over 10 distinct breeding years in a population where individuals experienced variability in food availability and habitat disruption. Litters of primiparous dams became increasingly female-biased with increasing litter size, but that trend waned among second litters born to dams, and reversed among third litters, with larger litters becoming more male-biased, suggesting that ASA is a product of interacting selection pressures. Trade-offs were not associated with habitat disruption, the availability of supplementary food, or dam age. An association between habitat disruption and male-biased sex ratios, the prevalence of litter size-sex ratio trade-offs and placental scar counts exceeding the number of juveniles at weaning in our population, but not in a geographically distinct population of conspecifics exposed to different environmental conditions reveal that the expression of ASA varies among populations and among years within populations, illustrating the conditional nature of ASA.

Empirical Overview of Benchmark Datasets for Geomagnetic Field-Based Indoor Positioning.

Indoor positioning and localization have been regarded as some of the most widely researched areas during the last decade. The wide proliferation of smartphones and the availability of fast-speed internet have initiated several location-based services. Concerning the importance of precise location information, many sensors are embedded into modern smartphones. Besides Wi-Fi positioning, a rich variety of technologies have been introduced or adopted for indoor positioning such as ultrawideband, infrared, radio frequency identification, Bluetooth beacons, pedestrian dead reckoning, and magnetic field, etc. However, special emphasis is put on infrastructureless approaches like Wi-Fi and magnetic field-based positioning, as they do not require additional infrastructure. Magnetic field positioning is an attractive solution for indoors; yet lack of public benchmarks and selection of suitable benchmarks are among the big challenges. While several benchmarks have been introduced over time, the selection criteria of a benchmark are not properly defined, which leads to positioning results that lack generalization. This study aims at analyzing various public benchmarks for magnetic field positioning and highlights their pros and cons for evaluation positioning algorithms. The concept of DUST (device, user, space, time) and DOWTS (dynamicity, orientation, walk, trajectory, and sensor fusion) is introduced which divides the characteristics of the magnetic field dataset into basic and advanced groups and discusses the publicly available datasets accordingly.

Association and response accuracy in the wild.

We studied contestant accuracy and error in a popular television quiz show, "Jeopardy!" Using vector-based knowledge representations obtained from distributional models of semantic memory, we computed the strength of association between clues and responses in over 5,000 televised games. Such representations have been shown to play a key role in memory and judgment, and consistent with this work, we find that contestants are more likely to provide correct responses when these responses are strongly associated with their clues, and more likely to provide incorrect responses when correct responses are weakly or negatively associated with their clues. This effect is stronger for easier questions with low monetary values, and for questions in which contestants compete to respond quickly. Our results show how distributional models of semantic memory can be used to predict human behavior in naturalistic high-level judgment tasks with skilled participants and significant monetary and social incentives.

Real-Time Massive Vector Field Data Processing in Edge Computing.

The spread of the sensors and industrial systems has fostered widespread real-time data processing applications. Massive vector field data (MVFD) are generated by vast distributed sensors and are characterized by high distribution, high velocity, and high volume. As a result, computing such kind of data on centralized cloud faces unprecedented challenges, especially on the processing delay due to the distance between the data source and the cloud. Taking advantages of data source proximity and vast distribution, edge computing is ideal for timely computing on MVFD. Therefore, we are motivated to propose an edge computing based MVFD processing framework. In particular, we notice that the high volume feature of MVFD results in high data transmission delay. To solve this problem, we invent Data Fluidization Schedule (DFS) in our framework to reduce the data block volume and the latency on Input/Output (I/O). We evaluated the efficiency of our framework in a practical application on massive wind field data processing for cyclone recognition. The high efficiency our framework was verified by the fact that it significantly outperformed classical big data processing frameworks Spark and MapReduce.

An Implantable Inductive Near-Field Communication System with 64 Channels for Acquisition of Gastrointestinal Bioelectrical Activity.

High-resolution (HR) mapping of the gastrointestinal (GI) bioelectrical activity is an emerging method to define the GI dysrhythmias such as gastroparesis and functional dyspepsia. Currently, there is no solution available to conduct HR mapping in long-term studies. We have developed an implantable 64-channel closed-loop near-field communication system for real-time monitoring of gastric electrical activity. The system is composed of an implantable unit (IU), a wearable unit (WU), and a stationary unit (SU) connected to a computer. Simultaneous data telemetry and power transfer between the IU and WU is carried out through a radio-frequency identification (RFID) link operating at 13.56 MHz. Data at the IU are encoded according to a self-clocking differential pulse position algorithm, and load shift keying modulated with only 6.25% duty cycle to be back scattered to the WU over the inductive path. The retrieved data at the WU are then either transmitted to the SU for real-time monitoring through an ISM-band RF transceiver or stored locally on a micro SD memory card. The measurement results demonstrated successful data communication at the rate of 125 kb/s when the distance between the IU and WU is less than 5 cm. The signals recorded in vitro at IU and received by SU were verified by a graphical user interface.

Calibration, Conversion, and Quantitative Multi-Layer Inversion of Multi-Coil Rigid-Boom Electromagnetic Induction Data.

Multi-coil electromagnetic induction (EMI) systems induce magnetic fields below and above the subsurface. The resulting magnetic field is measured at multiple coils increasingly separated from the transmitter in a rigid boom. This field relates to the subsurface apparent electrical conductivity (σa), and σa represents an average value for the depth range investigated with a specific coil separation and orientation. Multi-coil EMI data can be inverted to obtain layered bulk electrical conductivity models. However, above-ground stationary influences alter the signal and the inversion results can be unreliable. This study proposes an improved data processing chain, including EMI data calibration, conversion, and inversion. For the calibration of σa, three direct current resistivity techniques are compared: Electrical resistivity tomography with Dipole-Dipole and Schlumberger electrode arrays and vertical electrical soundings. All three methods obtained robust calibration results. The Dipole-Dipole-based calibration proved stable upon testing on different soil types. To further improve accuracy, we propose a non-linear exact EMI conversion to convert the magnetic field to σa. The complete processing workflow provides accurate and quantitative EMI data and the inversions reliable estimates of the intrinsic electrical conductivities. This improves the ability to combine EMI with, e.g., remote sensing, and the use of EMI for monitoring purposes.

On using wearable tri-axial accelerometers to examine the striking phase kinematics of expert specialist drag flickers on-field.

Analysing player kinematics during a match using "gold-standard" 3D video-based motion analysis techniques is a difficult prospect indeed. The development of small, wireless, wearable sensors offers the potential to reduce the challenges of measuring kinematics during match-play without hindering performance. The present study examined the viability of using wireless tri-axial accelerometers to examine whether key performance measures of drag flicks executed by expert specialist drag-flickers are predicted by the kinematics of the striking phase. Linear mixed models were used to examine whether the speed and accuracy of players' drag flicks were predicted by the duration of stick-ball contact, and the kinematics of the lead lower limb at stick-ball contact and ball release. Results revealed that stick and lead lower limb kinematics significantly predicted shot accuracy but not shot speed. Shorter drag-time predicted more accurate flicks (p = 0.03) as did a more vertical leg at stick-ball contact (p = 0.016) and a more horizontal thigh at ball release (p = 0.001). This may indicate that there are more ways to produce fast drag flicks than accurate ones. This study illustrates that wireless tri-axial accelerometers can be used on-field to measure the effects of kinematics on key performance measures.

Satellite Images Combined with Field Data Reveal Negative Changes in the Distribution of Babassu Palms after Clearing off Amazonian Forests.

When the Amazonian rain forest is cut to create pasture, some of the original vegetal species survive clearing, even expressing their ability to invade agro-systems. It is true of the babassu palm, which can be considered, paradoxically, a natural resource by the "Interstate Movement of Babassu Fruit Breaker Women" or as native weed by land owners-farmers. To manage potential conflict of land uses, we study here the current density of this palm tree in different habitats, based on a combination of field data and remote sensing data. Firstly, we checked that the field survey methodology (i.e., counting free-trunk palm trees over 20 cm in circumference) provides density values compatible with those stemming from satellite images interpretation. We can see then that, a PA-Benfica Brazilian territory revealed an average density of the babassu lower in pastures (2.86 ind/ha) than in the dense forest (4.72 ind/ha) from which they originate and than in fallow land (4.31 ind/ha). We analyze in detail density data repartition in three habitats and we discuss results from the literature on the density of this palm tree versus its resilience at different developmental stages after forest clearing, depending on anthropogenic-or not-factors, including solar radiation, fire, weeding, clear cutting, burying fruit, and competition with forage grass. All these results can be exploited for the design of future management plans for the babassu palm and we think that the linked methodology and interdisciplinary approach can be extended to others palms and trees species in similar problematic issues.

Cadmium Bioaccumulation in Aquatic Oligochaetes Using a Biodynamic Model: A Review of Values of Physiological Parameters and Model Validation Using Laboratory and Field Bioaccumulation Data.

This study reviews certain physiological digestive parameters in the literature that could be used to predict tissue residues in aquatic oligochaetes using the biodynamic model. Predictions were evaluated with independently measured Cd bioaccumulation data in sediment bioassays and field oligochaetes. The parameter review focused on three species commonly used in ecotoxicity testing and bioaccumulation studies: Tubifex tubifex (Tt), Limnodrilus hoffmeisteri (Lh) and Lumbriculus variegatus (Lv). Median Ingestion rates (g g-1 d-1, dw) at unpolluted conditions were 7.8 (Tt), 24.5 (Lh) and 11.5 (Lv), while results were lower (1.7-2.4) at polluted conditions. Assimilation efficiencies ranged from 3.4-19.6% (Tt), 2.7-16.1% (Lh), and 10.9-25.6% (Lv). The biodynamic model accurately predicted Cd tissue concentration in T. tubifex exposed to spiked sediments in laboratory bioassays. Comparisons of predicted vs. measured Cd tissue concentration in bioassays or field aquatic oligochaetes suggest that the biodynamic model can predict Cd tissue concentration within a factor of five in 81.3% of cases, across a range of measured tissue concentrations from 0.1 to 100 µg Cd g-1 dw. Predictions can be refined by using physiological parameter values that have been measured under varying environmental conditions (e.g. temperature, dissolved oxygen). The model can underestimate tissue concentration by up to one order of magnitude when worms are exposed to highly contaminated sediments. Contrarily, predictions overestimate tissue concentration by up to two orders of magnitude when the measured Cd < 0.1 µg g-1 dw, although in most cases these predictions do not fail bioaccumulation-based risk assessments, using a tissue threshold value of 1.5 µg Cd g-1 dw.

Improving the thermal efficiency of a jaggery production module using a fire-tube heat exchanger.

Jaggery is a product obtained after heating and evaporation processes have been applied to sugar cane juice via the addition of thermal energy, followed by the crystallisation process through mechanical agitation. At present, jaggery production uses furnaces and pans that are designed empirically based on trial and error procedures, which results in low ranges of thermal efficiency operation. To rectify these deficiencies, this study proposes the use of fire-tube pans to increase heat transfer from the flue gases to the sugar cane juice. With the aim of increasing the thermal efficiency of a jaggery installation, a computational fluid dynamic (CFD)-based model was used as a numerical tool to design a fire-tube pan that would replace the existing finned flat pan. For this purpose, the original configuration of the jaggery furnace was simulated via a pre-validated CFD model in order to calculate its current thermal performance. Then, the newly-designed fire-tube pan was virtually replaced in the jaggery furnace with the aim of numerically estimating the thermal performance at the same operating conditions. A comparison of both simulations highlighted the growth of the heat transfer rate at around 105% in the heating/evaporation processes when the fire-tube pan replaced the original finned flat pan. This enhancement impacted the jaggery production installation, whereby the thermal efficiency of the installation increased from 31.4% to 42.8%.

Influence of the spatial heterogeneity in tick abundance in the modeling of the seasonal activity of Ixodes ricinus nymphs in Western Europe.

The seasonal weather-driven activity of the tick Ixodes ricinus is frequently explored using multisite surveys. This study aimed to investigate the statistical modeling of seasonal trends in the activity of I. ricinus nymphs when both the influence of abiotic factors and spatial heterogeneity were taken into account. Time series data of abiotic covariates (temperature, relative humidity, rainfall and photoperiod) and nymphal tick counts were recorded on several sites in The Netherlands, Belgium and in France in 2008 and 2009. The sites were divided into two subsets which were used for model construction or model validation. A generalized linear mixed model was set up, with aggregated abiotic covariates as fixed effects, and the collection site as a random effect to account for the site-varying density in nymphs. A linear regression model was developed to estimate the site effect against the observed local abundance on each site. The activity patterns simulated from the weather and photoperiod covariates realistically reproduced the observed seasonal trends in nymphal tick activity. The fit between observed and simulated nymphal count time series was greatly improved when the site-specific local abundance in nymphs was included. Our modeling approach allows indicators of local tick abundance and the temporal modeling of I. ricinus activity to be combined. The model presented here can also be used to study scenarios on the temporal patterns of I. ricinus activity in the present and in the context of climate change.

Preliminary validation study of the 3-min wrist-worn psychomotor vigilance test.

The 10-min psychomotor vigilance test (PVT) is a widely used tool to assess behavioral alertness. In operational environments, however, the 3-min version of the PVT is more practicable. Under controlled laboratory conditions, we compared performance on the 3-min PVT on a wrist-worn device with the 3-min PVT on a laptop computer. We conducted two repeated measures experiments (72 participants) to assess the effects of PVT device type (laptop or wrist-worn), backlight on the wrist-worn device (on or off), ambient light (low or normal levels), and interstimulus interval (ISI). First, we compared the wrist-worn PVT (ISI of 2-10 s) with the laptop-based PVT (ISI of 1-4 s). Participants performed six PVT trials counterbalanced for order of ambient light and device type. In both ambient-lighting conditions, the median differences in PVT metrics (reaction time, response speed, and percentage of 355-ms or 500-ms lapses combined with false starts) between the laptop and the wrist-worn PVT with the backlight on were less than 4.5 %. Specifically, the median difference in reaction times was less than 10 ms. However, the standard deviation in most metrics using the wrist-worn PVT was twice that of the laptop. In a second experiment, we compared the wrist-worn and laptop PVTs with the same ISIs (1-4 s) in normal light conditions. The results suggest that PVT data can be collected reliably in the field using the wrist-worn device with the backlight on and an ISI of 2-10 s.

Area-based fuzzy membership forest cover comparison between MODIS NPP and Forest Inventory and Analysis (FIA) across eastern U.S. forest.

This research assessed the accuracy of the moderate resolution imaging spectroradiometer's (MODIS) land cover classification of softwood and hardwood using a fuzzy-based approach for 31 easternmost states in the U.S. Our main objective was to quantitatively evaluate spatially explicit land cover classifications of MODIS net primary product (NPP) scheme using the USDA Forest Service's (FS) field-based, tree-specific Forest Inventory Analysis (FIA). We used a grid of 648 km2 hexagons as base mapping units and interpreted our results at the USDA FS level IV ecological regions. Forest area was calculated for both MODIS and FIA and were found to be strongly correlated (Pearson's r = 0.875, p < 0.01), which suggests the two classifications are comparable. Area-based fuzzy memberships of softwood and hardwood forest were determined for both MODIS and FIA for each hexagon. We used cross-entropy (H c) to evaluate the accuracy of the MODIS classification. Our results determined that the accuracy of MODIS forest cover classification was not uniform for all ecological regions. Tree species importance values (IV) and Shannon's diversity index (H s) were calculated to examine species abundance and heterogeneity, which may partially explain discrepancies between MODIS and FIA classifications. The greatest misclassifications were due to (1) MODIS underestimating softwood forest cover and (2) MODIS confusing forest cover with other land covers such as grassland, cropland, or woody savanna. Our results provide a guideline for users to understand the degree of uncertainty of MODIS forest cover classifications in the eastern USA.

Contribution of aqueous and dietary uptakes to lead (Pb) bioaccumulation in Gammarus pulex: From multipathway modeling to in situ validation.

Although dynamic approaches are nowadays used increasingly to describe metal bioaccumulation in aquatic organisms, the validation of such laboratory-derived modeling is rarely assessed under environmental conditions. Furthermore, information on bioaccumulation kinetics of Pb and the significance of its uptake by dietary route is scarce in freshwater species. This study aims at modeling aqueous and dietary uptakes of Pb in the litter-degrader Gammarus pulex and assessing the predictive quality of multipathway modeling from in situ bioaccumulation data. In microcosms, G. pulex were exposed to environmentally realistic concentrations of Pb (from 0.1 to 10µg/L) in the presence of Pb-contaminated poplar leaves, which were enclosed or not in a net to distinguish aqueous and dietary uptakes. Results show that water and food both constitute contamination sources for gammarids. Establishing biodynamic parameters involved in Pb aqueous and dietary uptake and elimination rates enabled to construct a multipathway model to describe Pb bioaccumulation in gammarids. This laboratory-derived model successfully predicted bioaccumulation measured in native populations of G. pulex collected in situ when local litter was used as dietary exposure source. This study demonstrates not only the suitable applicability of biodynamic parameters for predicting Pb bioaccumulation but also the necessity of taking dietary uptake into account for a better interpretation of the gammarids' contamination in natural conditions.

Evaluation of two pesticide leaching models in an irrigated field cropped with corn.

Pesticide leaching models is an easy and cost effective method used in the prediction of surface and groundwater pollution. In this paper, the ability of two pesticide leaching models, MACRO and PEARL, to describe soil water dynamics and atrazine's transport through the soil profile was examined. The data used for the comparison was obtained from an experiment in an irrigated corn field in the plain of the Ardas River, in north-eastern Greece. Both models were parameterized using pedotransfer functions, field and laboratory data. The uncalibrated simulation showed several discrepancies, therefore the retention curve and the sorption parameters were calibrated according to the trial and error method. The comparison of both models indicated that soil water flow was described similarly. The simulated results of atrazine's concentration were evaluated and compared to the measured concentrations at specific depths, using statistical criteria. Atrazine transport was simulated in a satisfactory manner as confirmed by model efficiency (EF) values, that are very close to unit. Coefficient of residual mass (CRM) values for both models are positive, indicating that both models underestimate the measured data. MACRO estimated higher accumulated actual evapotranspiration values, and less percolated water from soil profile than PEARL, and as a result, change in water content was higher in the latter. PEARL also predicted that half the amount of the applied mass was decayed two days earlier than the day estimated by MACRO. Generally, MACRO simulated the fate of atrazine in soil better than PEARL.

Healthy, wealthy, and wise: retirement planning predicts employee health improvements.

Are poor physical and financial health driven by the same underlying psychological factors? We found that the decision to contribute to a 401(k) retirement plan predicted whether an individual acted to correct poor physical-health indicators revealed during an employer-sponsored health examination. Using this examination as a quasi-exogenous shock to employees' personal-health knowledge, we examined which employees were more likely to improve their health, controlling for differences in initial health, demographics, job type, and income. We found that existing retirement-contribution patterns and future health improvements were highly correlated. Employees who saved for the future by contributing to a 401(k) showed improvements in their abnormal blood-test results and health behaviors approximately 27% more often than noncontributors did. These findings are consistent with an underlying individual time-discounting trait that is both difficult to change and domain interdependent, and that predicts long-term individual behaviors in multiple dimensions.

A comprehensive study of settlement during the filling and post-closure phases at a landfill in Québec, Canada: Field data and TMB modelling.

In Northern climates, waste placed curbside the evening before waste collection can lead to partially frozen waste at placement, which delays biodegradation and biodegradation-induced settlement. A 12-year settlement dataset collected during the filling and post-closure phases at a landfill in Québec, Canada was analyzed. The dataset showed a delay in biodegradation-induced settlement due to the first three waste lifts being placed in the winter months and exhibited an increase in the settlement rate at later times when the waste temperatures increased to values that support biodegradation. The field data also demonstrated that the stiffness of MSW increased in response to confined stress as subsequent waste lifts were added. A thermal-mechanical-biological (TMB) model was developed, in COMSOL Multiphysics, to simulate the settlement dataset. TMB integrates a Generalized Kelvin-Voigt (GKV) model, simulating instantaneous and mechanical creep settlements, with a biodegradation-induced settlement model that relates heat/gas generation with time to biodegradation-induced settlement. The thermal model simulates heat transfer through conduction and includes a biodegradation heat generation source term. The GKV stiffness parameters are expressed as a function of the applied stress to account for waste compressibility effects on mechanical response, which is consistent with field data and the research literature. The paper focuses solely on the MSW settlement field data and model predictions, with thermal response analysis presented in a separate publication. The TMB model effectively predicted waste behaviour, including resistance to compressibility under higher stress and the delay in waste settlement for waste placed in winter. The temperature and settlement data provide a valuable dataset to validate different models that can be used to predict waste settlement in cold regions.