scholarly journals Machine Learning Approach to Simulate Soil CO2 Fluxes under Cropping Systems

Agronomy ◽  
2022 ◽  
Vol 12 (1) ◽  
pp. 197
Author(s):  
Toby A. Adjuik ◽  
Sarah C. Davis
Keyword(s):  
Machine Learning ◽  
Soil Temperature ◽  
Greenhouse Gas ◽  
Air Temperature ◽  
Ghg Emissions ◽  
Field Measurements ◽  
Soil Classification ◽  
Support Vector ◽  
Co2 Fluxes ◽  
Soil Co2

With the growing number of datasets to describe greenhouse gas (GHG) emissions, there is an opportunity to develop novel predictive models that require neither the expense nor time required to make direct field measurements. This study evaluates the potential for machine learning (ML) approaches to predict soil GHG emissions without the biogeochemical expertise that is required to use many current models for simulating soil GHGs. There are ample data from field measurements now publicly available to test new modeling approaches. The objective of this paper was to develop and evaluate machine learning (ML) models using field data (soil temperature, soil moisture, soil classification, crop type, fertilization type, and air temperature) available in the Greenhouse gas Reduction through Agricultural Carbon Enhancement network (GRACEnet) database to simulate soil CO2 fluxes with different fertilization methods. Four machine learning algorithms—K nearest neighbor regression (KNN), support vector regression (SVR), random forest (RF) regression, and gradient boosted (GB) regression—were used to develop the models. The GB regression model outperformed all the other models on the training dataset with R2 = 0.88, MAE = 2177.89 g C ha−1 day−1, and RMSE 4405.43 g C ha−1 day−1. However, the RF and GB regression models both performed optimally on the unseen test dataset with R2 = 0.82. Machine learning tools were useful for developing predictors based on soil classification, soil temperature and air temperature when a large database like GRACEnet is available, but these were not highly predictive variables in correlation analysis. This study demonstrates the suitability of using tree-based ML algorithms for predictive modeling of CO2 fluxes, but no biogeochemical processes can be described with such models.

Modeling soil temperature using air temperature features in diverse climatic conditions with complementary machine learning models

2021 ◽  
Vol 185 ◽  
pp. 106158
Author(s):  
Maryam Bayatvarkeshi ◽  
Suraj Kumar Bhagat ◽  
Kourosh Mohammadi ◽  
Ozgur Kisi ◽  
M. Farahani ◽  
...  
Keyword(s):  
Machine Learning ◽  
Soil Temperature ◽  
Air Temperature ◽  
Climatic Conditions ◽  
Learning Models ◽  
Machine Learning Models

Remotely sensed temperature is a proxy of greenhouse gas emissions in intact and managed peatlands

2021 ◽  
Author(s):  
Ain Kull ◽  
Iuliia Burdun ◽  
Gert Veber ◽  
Oleksandr Karasov ◽  
Martin Maddison ◽  
...  
Keyword(s):  
Soil Temperature ◽  
Greenhouse Gas Emissions ◽  
Greenhouse Gas ◽  
Water Table ◽  
Ecosystem Respiration ◽  
Ghg Emissions ◽  
Remotely Sensed ◽  
Water Table Depth ◽  
Gas Emissions

<p>Besides water table depth, soil temperature is one of the main drivers of greenhouse gas (GHG) emissions in intact and managed peatlands. In this work, we evaluate the performance of remotely sensed land surface temperature (LST) as a proxy of greenhouse gas emissions in intact, drained and extracted peatlands. For this, we used chamber-measured carbon dioxide (CO<sub>2</sub>) and methane (CH<sub>4</sub>) data from seven peatlands in Estonia collected during vegetation season in 2017–2020. Additionally, we used temperature and water table depth data measured in situ. We studied relationships between CO<sub>2</sub>, CH<sub>4</sub>, in-situ parameters and remotely sensed LST from Landsat 7 and 8, and MODIS Terra. Results of our study suggest that LST has stronger relationships with surface and soil temperature as well as with ecosystem respiration (R<sub>eco</sub>) over drained and extracted sites than over intact ones. Over the extracted cites the correlation between R<sub>eco</sub> CO<sub>2</sub> and LST is 0.7, and over the drained sites correlation is 0.5. In natural sites, we revealed a moderate positive relationship between LST and CO<sub>2</sub> emitted in hollows (correlation is 0.6) while it is weak in hummocks (correlation is 0.3). Our study contributes to the better understanding of relationships between greenhouse gas emissions and their remotely sensed proxies over peatlands with different management status and enables better spatial assessment of GHG emissions in drainage affected northern temperate peatlands.</p>

scholarly journals Synthesizing greenhouse gas fluxes across nine European peatlands and shrublands – responses to climatic and environmental changes

2012 ◽  
Vol 9 (3) ◽  
pp. 3693-3738 ◽  
Author(s):  
M. S. Carter ◽  
K. S. Larsen ◽  
B. Emmett ◽  
M. Estiarte ◽  
C. Field ◽  
...  
Keyword(s):  
Global Warming ◽  
Greenhouse Gas ◽  
Air Temperature ◽  
Environmental Changes ◽  
Negative Relationship ◽  
Co2 Efflux ◽  
Soil Co2 ◽  
Prolonged Drought ◽  
Mean Annual Air Temperature

Abstract. In this study, we compare annual fluxes of methane (CH4), nitrous oxide (N2O) and soil respiratory carbon dioxide (CO2) measured at nine European peatlands (n = 4) and shrublands (n = 5). The sites range from northern Sweden to Spain, covering a span in mean annual air temperature from 0 to 16 °C, and in annual precipitation from 300 to 1300 mm yr−1. The effects of climate change, including temperature increase and prolonged drought, were tested at five shrubland sites. At one peatland site, the long-term (>30 yr) effect of drainage was assessed, while increased nitrogen deposition was investigated at three peatland sites. The shrublands were generally sinks for atmospheric CH4 whereas the peatlands were CH4 sources, with fluxes ranging from −519 to +6890 mg CH4-C m−2 yr−1 across the studied ecosystems. At the peatland sites, annual CH4 emission increased with mean annual air temperature, while a negative relationship was found between net CH4 uptake and the soil carbon stock at the shrubland sites. Annual N2O fluxes were generally small ranging from –14 to 42 mg N2O-N m−2 yr−1. Highest N2O emission occurred at the sites that had highest concentration of nitrate (NO3−) in soil water. Furthermore, experimentally increased NO3− deposition led to increased N2O efflux, whereas prolonged drought and long-term drainage reduced the N2O efflux. Soil CO2 emissions in control plots ranged from 310 to 732 g CO2-C m−2 yr−1. Drought and long-term drainage generally reduced the soil CO2 efflux, except at a~hydric shrubland where drought tended to increase soil respiration. When comparing the fractional importance of each greenhouse gas to the total numerical global warming response, the change in CO2 efflux dominated the response in all treatments (ranging 71–96%), except for NO3− addition where 89% was due to change in CH4 emissions. Thus, in European peatlands and shrublands the feedback to global warming induced by the investigated anthropogenic disturbances will be dominated by variations in soil CO2 fluxes.

Comparison of RZWQM2 and DNDC Models to Simulate Greenhouse Gas Emissions under Combined Inorganic/Organic Fertilization in a Subsurface-Drained Field

2020 ◽  
Vol 63 (4) ◽  
pp. 771-787
Author(s):  
Qianjing Jiang ◽  
Zhiming Qi ◽  
Chandra A. Madramootoo ◽  
Ward Smith ◽  
Naeem A. Abbasi ◽  
...  
Keyword(s):  
Soil Temperature ◽  
Greenhouse Gas Emissions ◽  
Greenhouse Gas ◽  
Water Content ◽  
Soil Water Content ◽  
Ghg Emissions ◽  
Measured Data ◽  
Gas Emissions ◽  
N Management ◽  
Better Than

HighlightsRZWQM2 was compared with DNDC to predict greenhouse gas emissions.RZWQM2 was applied to simulate the greenhouse gas emissions under manure application.RZWQM2 performed better than DNDC in simulating soil water content and CO2 emissions.Abstract. N management has the potential to mitigate greenhouse gas (GHG) emissions. Process-based models are promising tools for evaluating and developing management practices that may optimize sustainability goals as well as promote crop productivity. In this study, the GHG emission component of the Root Zone Water Quality Model (RZWQM2) was tested under two different types of N management and subsequently compared with the Denitrification-Decomposition (DNDC) model using measured data from a subsurface-drained field with a corn-soybean rotation in southern Ontario, Canada. Field-measured data included N2O and CO2 fluxes, soil temperature, and soil moisture content from a four-year field experiment (2012 to 2015). The experiment was composed of two N treatments: inorganic fertilizer (IF), and inorganic fertilizer combined with solid cattle manure (SCM). Both models were calibrated using the data from IF and validated with SCM. Statistical results indicated that both models predicted well the soil temperature, but RZWQM2 performed better than DNDC in simulating soil water content (SWC) because DNDC lacked a heterogeneous soil profile, had shallow simulation depth, and lacked crop root density functions. Both RZWQM2 and DNDC predicted the cumulative N2O and CO2 emissions within 15% error under all treatments, while the timing of daily CO2 emissions was more accurately predicted by RZWQM2 (RMSE = 0.43 to 0.54) than by DNDC (RMSE = 0.60 to 0.67). Modeling results for N management effects on GHG emissions showed consistency with the field measurements, indicating higher CO2 emissions under SCM than IF, higher N2O emissions under IF in corn years, but lower N2O emissions in soybean years. Overall, RZWQM2 required more experienced and intensive calibration and validation, but it provided more accurate predictions of soil hydrology and better timing of CO2 emissions than DNDC. Keywords: CO2 emission, Corn-soybean rotation, Inorganic fertilization, Manure application, N2O emission, Process-based modeling.

A Cluster Analysis Study of Opportune Adoption of Electric Drive Vehicles for Better Greenhouse Gas Reduction

2016 ◽  
Author(s):  
Karim Hamza ◽  
Kenneth P. Laberteaux
Keyword(s):  
Greenhouse Gas ◽  
Electric Drive ◽  
Ghg Emissions ◽  
Support Vector ◽  
Driving Behaviors ◽  
Greenhouse Gas Reduction ◽  
Stop And Go ◽  
Ghg Reduction ◽  
Vector Clustering ◽  
Household Travel

Adoption of electric drive vehicles (EDVs) presents an opportunity for reduction of greenhouse gas (GHG) emissions. From an individual vehicle standpoint however, the GHG reduction can vary significantly depending on the type of driving that the vehicle is used for. This is primarily due to conventional vehicles (CVs) having poor energy efficiency in stop-and-go city-like driving compared to their performance in steady highway-like driving. This study attempts to examine the magnitude of the differential in GHG reduction benefit for real driving behaviors obtained from California Household Travel Survey (CHTS-2013). Recorded vehicles speed traces are analyzed via a fuel economy simulator then a hybrid support vector clustering (SVC) technique is applied to form groups of vehicle samples with similar driving behaviors. Unlike many clustering techniques, SVC does not impose a pre-dictated number of clusters, but has a number of parameters that must be tuned in order to obtain meaningful results. Tuning of the parameters is performed via a multi-objective evolutionary algorithm (SPEA2) after formulating the cluster tuning as a two-objective problem that seeks to maximize: i) differential benefit in GHG reduction, and ii) fraction of the population that groups of vehicles represent. Results show that replacing a CV with its equivalent hybrid (HEV) can reduce GHG emissions per mile of driving by 2 to 2.5 times more for a group of vehicles (best opportune for an EDV) compared to the less opportune group.

scholarly journals Development of a Decision Support Model Based on Machine Learning for Applying Greenhouse Gas Reduction Technology

2020 ◽  
Vol 12 (9) ◽  
pp. 3582
Author(s):  
Sungwoo Lee ◽  
Sungho Tae
Keyword(s):  
Machine Learning ◽  
Renewable Energy ◽  
Greenhouse Gas ◽  
Ghg Emissions ◽  
Assessment Method ◽  
Economic Benefits ◽  
Environmental Benefits ◽  
Assessment Model ◽  
Decision Support Model ◽  
Ghg Reduction

Multiple nations have implemented policies for greenhouse gas (GHG) reduction since the 21st Conference of Parties (COP 21) at the United Nations Framework Convention on Climate Change (UNFCCC) in 2015. In this convention, participants voluntarily agreed to a new climate regime that aimed to decrease GHG emissions. Subsequently, a reduction in GHG emissions with specific reduction technologies (renewable energy) to decrease energy consumption has become a necessity and not a choice. With the launch of the Korean Emissions Trading Scheme (K-ETS) in 2015, Korea has certified and financed GHG reduction projects to decrease emissions. To help the user make informed decisions for economic and environmental benefits from the use of renewable energy, an assessment model was developed. This study establishes a simple assessment method (SAM), an assessment database (DB) of 1199 GHG reduction technologies implemented in Korea, and a machine learning-based GHG reduction technology assessment model (GRTM). Additionally, we make suggestions on how to evaluate economic benefits, which can be obtained in conjunction with the environmental benefits of GHG reduction technology. Finally, we validate the applicability of the assessment model on a public building in Korea.

Experiments on Machine Learning Techniques for Soil Classification Using Sentinel-2 Products

2020 ◽  
Author(s):  
Victor Bacu ◽  
Teodor Stefanut ◽  
Dorian Gorgan
Keyword(s):  
Machine Learning ◽  
High Resolution ◽  
Satellite Data ◽  
Field Measurements ◽  
Soil Classification ◽  
Machine Learning Techniques ◽  
Classification Models ◽  
Spatial Features ◽  
Learning Techniques ◽  
Sentinel 2

<p>Agricultural management relies on good, comprehensive and reliable information on the environment and, in particular, the characteristics of the soil. The soil composition, humidity and temperature can fluctuate over time, leading to migration of plant crops, changes in the schedule of agricultural work, and the treatment of soil by chemicals. Various techniques are used to monitor soil conditions and agricultural activities but most of them are based on field measurements. Satellite data opens up a wide range of solutions based on higher resolution images (i.e. spatial, spectral and temporal resolution). Due to this high resolution, satellite data requires powerful computing resources and complex algorithms. The need for up-to-date and high-resolution soil maps and direct access to this information in a versatile and convenient manner is essential for pedology and agriculture experts, farmers and soil monitoring organizations.</p><p>Unfortunately, the satellite image processing and interpretation are very particular to each area, time and season, and must be calibrated by the real field measurements that are collected periodically. In order to obtain a fairly good accuracy of soil classification at a very high resolution, without using interpolation methods of an insufficient number of measurements, the prediction based on artificial intelligence techniques could be used. The use of machine learning techniques is still largely unexplored, and one of the major challenges is the scalability of the soil classification models toward three main directions: (a) adding new spatial features (i.e. satellite wavelength bands, geospatial parameters, spatial features); (b) scaling from local to global geographical areas; (c) temporal complementarity (i.e. build up the soil description by samples of satellite data acquired along the time, on spring, on summer, in another year, etc.).</p><p>The presentation analysis some experiments and highlights the main issues on developing a soil classification model based on Sentinel-2 satellite data, machine learning techniques and high-performance computing infrastructures. The experiments concern mainly on the features and temporal scalability of the soil classification models. The research is carried out using the HORUS platform [1] and the HorusApp application [2], [3], which allows experts to scale the computation over cloud infrastructure.</p><p> </p><p>References:</p><p>[1] Gorgan D., Rusu T., Bacu V., Stefanut T., Nandra N., “Soil Classification Techniques in Transylvania Area Based on Satellite Data”. World Soils 2019 Conference, 2 - 3 July 2019, ESA-ESRIN, Frascati, Italy (2019).</p><p>[2] Bacu V., Stefanut T., Gorgan D., “Building soil classification maps using HorusApp and Sentinel-2 Products”, Proceedings of the Intelligent Computer Communication and Processing Conference – ICCP, in IEEE press (2019).</p><p>[3] Bacu V., Stefanut T., Nandra N., Rusu T., Gorgan D., “Soil classification based on Sentinel-2 Products using HorusApp application”, Geophysical Research Abstracts, Vol. 21, EGU2019-15746, 2019, EGU General Assembly (2019).</p>

First field estimation of greenhouse gas emissions from European soil-dwelling Scarabaeidae larvae targeting the genus Melolontha

2020 ◽  
Author(s):  
Carolyn-Monika Görres ◽  
Claudia Kammann
Keyword(s):  
Greenhouse Gas ◽  
Carbon Emissions ◽  
Developmental Stages ◽  
Ghg Emissions ◽  
Field Measurements ◽  
Spatial And Temporal Variability ◽  
Activity Levels ◽  
Ghg Emission ◽  
Forest Sites ◽  
Wide Range

<p>Arthropods are a major soil fauna group, and have the potential to substantially influence the spatial and temporal variability of soil greenhouse gas (GHG) sinks and sources. The overall effect of soil-inhabiting arthropods on soil GHG fluxes still remains poorly quantified since the majority of the available data comes from laboratory experiments, is often controversial, and has been limited to a few species. The main objective of this study was to provide first insights into field-level carbon dioxide (CO<sub>2</sub>), methane (CH<sub>4</sub>) and nitrous oxide (N<sub>2</sub>O) emissions of soil-inhabiting larvae of the Scarabaeidae family. Larvae of the genus <em>Melolontha</em> were excavated at various grassland and forest sites in west-central and southern Germany, covering a wide range of different larval developmental stages, and larval activity levels. Excavated larvae were immediately incubated in the field to measure their GHG emissions. Gaseous carbon emissions of individual larvae showed a large inter- and intra-site variability which was strongly correlated to larval biomass. This correlation persisted when upscaling CO<sub>2</sub> and CH<sub>4 </sub>emissions to the plot scale. Field emission estimates for <em>Melolontha</em> spp. were subsequently upscaled to the European level to derive the first regional GHG emission estimates for members of the Scarabaeidae family. Estimates ranged between 10.42 and 409.53 kt CO<sub>2</sub> yr<sup>-1</sup>, and 0.01 and 1.36 kt CH<sub>4</sub> yr<sup>-1</sup>. Larval N<sub>2</sub>O emissions were only sporadically observed and not upscaled. For one site, a comparison of field- and laboratory-based GHG emission measurements was conducted to assess potential biases introduced by transferring Scarabaeidae larvae to artificial environments. Emission strength and variability of captive larvae decreased significantly within two weeks and the correlation between larval biomass and gaseous carbon emissions disappeared, highlighting the importance of field measurements. Overall, our data show that Scarabaeidae larvae can be significant soil GHG sources and should not be neglected in soil GHG flux research.</p>

scholarly journals Evaluating essential features of proppant transport at engineering scales combining field measurements with machine learning algorithms

2021 ◽  
Author(s):  
Xiaoyu Wang ◽  
Lei Hou ◽  
Xueyu Geng ◽  
Peibin Gong ◽  
Honglei Liu
Keyword(s):  
Machine Learning ◽  
Mean Squared Error ◽  
Learning Algorithms ◽  
Field Measurements ◽  
Stratified Flow ◽  
Machine Learning Algorithms ◽  
Support Vector ◽  
Proppant Transport ◽  
Control Variate ◽  
Injection Process

The characterization of the proppant transport at a field-engineering scale is still challenging due to the lack of direct subsurface measurements. Features that control the proppant transport may link the experimental and numerical observations to the practical operations at a field scale. To improve the numerical and laboratory simulations, we propose a machine-learning-based workflow to evaluate the essential features of proppant transport and their corresponding calculations. The proppant flow in fractures is estimated by applying the Gated recurrent unit (GRU) and Support-vector machine (SVM) algorithms to the measurements obtained from shale gas fracturing operations. Over 430,000 groups of fracturing data are collected and pre-processed by the proppant transport models to calculate key features, including settlement, stratified flow and inception of settled particles. The features are then fed into machine learning algorithms for pressure prediction. The root mean squared error (RMSE) is used as the criterion for ranking selected features via the control variate method. Our result shows that the stratified-flow feature (fracture-level) possesses better interpretations for the proppant transport, in which the Bi-power model helps to produce the best predictions. The settlement and inception features (particle-level) perform better in cases that the pressure fluctuates significantly, indicating that more complex fractures may have been generated. Moreover, our analyses on the remaining errors in the pressure-ascending cases suggest that (1) an introduction of the alternate-injection process, and (2) the improved calculation of proppant transport in complex fracture networks and highly-filled fractures will be beneficial to both experimental observations and field applications.

Sign in / Sign up

Export Citation Format

Share Document