Data

National surface and near-surface conductivity grids

Geoscience Australia
Wilford, J. ; LeyCooper, Y. ; Sudipta Basak
Viewed: [[ro.stat.viewed]] Cited: [[ro.stat.cited]] Accessed: [[ro.stat.accessed]]
ctx_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rfr_id=info%3Asid%2FANDS&rft_id=https://pid.geoscience.gov.au/dataset/ga/148588&rft.title=National surface and near-surface conductivity grids&rft.identifier=https://pid.geoscience.gov.au/dataset/ga/148588&rft.publisher=Commonwealth of Australia (Geoscience Australia)&rft.description=A national compilation of airborne electromagnetic (AEM) conductivity–depth models from AusAEM (Ley-Cooper et al. 2020) survey line data and other surveys (see reference list in the attachments) has been used to train a conductivity model prediction for the 0-4 m and 30 m depth intervals. Over 460,000 training points/measurements were used in a 5 K-Fold training and validation split. A further 28,626 points/measurements were used to assess the out of sample performance (OOS; i.e. points not used in the model validation). Modelling of the conductivity values (i.e. measurements along the AEM survey lines) was performed using the gradient boosted (GB) tree algorithm. The GB model is a machine learning (ML) ensemble technique used for both regression and classification tasks (https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingRegressor.html). Samples along the flight-line were thinned to approximately one sample per 300 m. This avoided the situation where we could have more than one sample per pixel (i.e. features or covariates used in the model prediction have a cell or pixel size of 80 m) that could otherwise lead to over fitting. In addition, out of sample set used label clusters or groups to minimise overfitting. Here we use the median of the models as the conductivity prediction and the upper and lower percentiles (95th and 5th respectively) to measure the model uncertainty. Grids show conductivity (S/m) in log 10 units. The methodology used to generate these conductivity grids are overall similar to that described by Wilford, et al. 2022. Reported out-of-sample r-squares for the 0-4 m and 3 m depths are 0.76 and 0.74, respectively. The ML approach allows estimation of conductivity into areas where we do not have airborne electromagnetic survey coverage. Hence these model have a national extent. Where we do not have AEM survey coverage the model is finding relationships with the covariates and making informed estimates of conductivity in those areas. Where those relationships are not well understood (i.e. where we see a departure in the feature space characteristics from what the model can ‘see’) the model prediction is likely to be less certain. Differences in the features and their corresponding values ‘seen’ and used in the model versus the full feature space covering the entire continent are captured in the covariate shift map. High values in the shift model can indicate higher potential uncertainty or unreliability of the model prediction. Users therefore need to be mindful when interpreting this dataset, of the uncertainties shown by the 5th-95th percentiles, and high values in the covariate shift map. Datasets in this data package include: 1. 0_4m_conductivity_prediction_median.tif2. 0_4m_conductivity_lower_percentile_5th.tif3. 0_4m_conductivity_upper_percentile_95th.tif4. 30m_conductivity_prediction_median.tif5.30m_conductivity_lower_percentile_5th.tif6. 30m_conductivity_upper_percentile_95th.tif7. National_conductivity_model_shift.tif8. Full list of referenced AEM survey datasets used to train the model (word document)9. Map showing the distribution of training and out-of-sample sitesAll the Geotiffs (1-6) are in log (10) electrical conductivity siemens per metre (S/m). This work is part of Geoscience Australia’s Exploring for the Future program which provides precompetitive information to inform decision-making by government, community and industry on the sustainable development of Australia's mineral, energy and groundwater resources. By gathering, analysing and interpreting new and existing precompetitive geoscience data and knowledge, we are building a national picture of Australia’s geology and resource potential. This leads to a strong economy, resilient society and sustainable environment for the benefit of all Australians. This includes supporting Australia’s transition to net zero emissions, strong, sustainable resources and agriculture sectors, and economic opportunities and social benefits for Australia’s regional and remote communities. The Exploring for the Future program, which commenced in 2016, is an eight year, $225m investment by the Australian Government.Reference:Ley-Cooper, A. Y., Brodie, R.C., and Richardson, M. 2020. AusAEM: Australia’s airborne electromagnetic continental-scale acquisition program, Exploration Geophysics, 51:1, 193-202, DOI: 10.1080/08123985.2019.1694393Wilford, J., LeyCooper, Y., Basak, S., Czarnota, K. 2022. High resolution conductivity mapping using regional AEM survey and machine learning. Geoscience Australia, Canberra. https://dx.doi.org/10.26186/146380Maintenance and Update Frequency: asNeededStatement: The conductivity grids are derived from a covariate machine learning approach that establishers predictive correlations between the AEM conductivity depth estimates with a suite of environmental and geological grids or covariates. The conductivity depth estimates are largely derived from the regional (20 km line-spaced) AusAEM survey data (Ley-Cooper et al. 2020). Conductivity-depth values were derived from a single model using Geoscience Australia's deterministic 1D smooth-30-layer layered-earth-inversion algorithm (Brodie and Richardson 2015). The covariates include terrain attributes, gamma radiometric, satellite imagery (including Barest Earth; Wilford J. and Roberts D., 2020), gravity and magnetic derivatives, geology polygons (1:1 million surface geology) and climate surfaces. The gradient boosted tree algorithm was used to generate the prediction models. Over 460,000 training points/measurements were used in the 5-Fold cross-validation and training the model. A further 28,626 points/measurements were kept aside to assess the out of sample performance (OOS; i.e. points not used in the model training and validation). Model performance was assessed using the OOS. The OOS used label clusters or groups to ensure spatial separation from sites used in the model training. The median of the models provide the conductivity prediction and the upper and lower percentiles (95th and 5th respectively) provides a measure model uncertainty. Grids show conductivity (S/m) in log 10 units. The overall modelling approach is described in Wilford et al., 2022.Ley-Cooper A. Y. & Brodie R. C., 2020. AusAEM: Imaging the near surface from the World’s Largest Airborne Electromagnetic Survey. In: Czarnota K., et al. (eds.) Exploring for the Future: Extended Abstracts, Geoscience Australia.http://dx.doi.org/10.11636/134528. Ross C Brodie & Murray Richardson (2015) Open Source Software for 1D Airborne Electromagnetic Inversion, ASEG Extended Abstracts, 2015:1, 1-3, DOI: 10.1071/ ASEG2015ab197 Wilford, J., LeyCooper, Y., Basak, S., Czarnota, K. 2022. High resolution conductivity mapping using regional AEM survey and machine learning. Geoscience Australia, Canberra. https://dx.doi.org/10.26186/146380Wilford, J. & Roberts D., 2020. Enhanced bare earth Landsat imagery for soil and lithological modelling. In: Czarnota K.,et al. (eds.), Exploring for the Future: extended abstracts, Geoscience Australia, Canberra,http://dx.doi.org/10.11636/134472.&rft.creator=Wilford, J. &rft.creator=LeyCooper, Y. &rft.creator=Sudipta Basak &rft.date=2023&rft.coverage=westlimit=112; southlimit=-44; eastlimit=154; northlimit=-9; projection=GDA94 / geographic 2D (EPSG: 4283)&rft.coverage=westlimit=112; southlimit=-44; eastlimit=154; northlimit=-9; projection=GDA94 / geographic 2D (EPSG: 4283)&rft_rights= https://creativecommons.org/licenses/by/4.0/&rft_rights=Creative Commons Attribution 4.0 International Licence&rft_rights=CC-BY&rft_rights=4.0&rft_rights=Any&rft_rights=Any&rft_rights=© Commonwealth of Australia (Geoscience Australia) 2023&rft_rights=Australian Government Security Classification System&rft_rights=https://www.protectivesecurity.gov.au/Pages/default.aspx&rft_rights=WWW:LINK-1.0-http--link&rft_rights=Australian Government Security Classification System&rft_rights=Creative Commons Attribution 4.0 International Licence http://creativecommons.org/licenses/by/4.0&rft_subject=geoscientificInformation&rft_subject=EFTF – Exploring for the Future&rft_subject=Australia's Resources Framework&rft_subject=conductivity&rft_subject=AusAEM interpretation&rft_subject=models&rft_subject=machine learning&rft_subject=australia&rft_subject=airbrone electromagnetic&rft_subject=Electrical and Electromagnetic Methods in Geophysics&rft_subject=Soil Physics&rft_subject=Published_External&rft.type=dataset&rft.language=English Access the data

Licence & Rights:

Open Licence view details
CC-BY

https://creativecommons.org/licenses/by/4.0/

Creative Commons Attribution 4.0 International Licence
http://creativecommons.org/licenses/by/4.0

Creative Commons Attribution 4.0 International Licence

CC-BY

4.0

Any

Any

© Commonwealth of Australia (Geoscience Australia) 2023

Australian Government Security Classification System

https://www.protectivesecurity.gov.au/Pages/default.aspx

WWW:LINK-1.0-http--link

Australian Government Security Classification System

Access:

Open

Brief description

A national compilation of airborne electromagnetic (AEM) conductivity–depth models from AusAEM (Ley-Cooper et al. 2020) survey line data and other surveys (see reference list in the attachments) has been used to train a conductivity model prediction for the 0-4 m and 30 m depth intervals. Over 460,000 training points/measurements were used in a 5 K-Fold training and validation split. A further 28,626 points/measurements were used to assess the out of sample performance (OOS; i.e. points not used in the model validation). Modelling of the conductivity values (i.e. measurements along the AEM survey lines) was performed using the gradient boosted (GB) tree algorithm. The GB model is a machine learning (ML) ensemble technique used for both regression and classification tasks (https://scikit-learn.org/stable/modules/generated/sklearn.ensemble.GradientBoostingRegressor.html). Samples along the flight-line were thinned to approximately one sample per 300 m. This avoided the situation where we could have more than one sample per pixel (i.e. features or covariates used in the model prediction have a cell or pixel size of 80 m) that could otherwise lead to over fitting. In addition, out of sample set used label clusters or groups to minimise overfitting. Here we use the median of the models as the conductivity prediction and the upper and lower percentiles (95th and 5th respectively) to measure the model uncertainty. Grids show conductivity (S/m) in log 10 units. The methodology used to generate these conductivity grids are overall similar to that described by Wilford, et al. 2022.
 
Reported out-of-sample r-squares for the 0-4 m and 3 m depths are 0.76 and 0.74, respectively. The ML approach allows estimation of conductivity into areas where we do not have airborne electromagnetic survey coverage. Hence these model have a national extent. Where we do not have AEM survey coverage the model is finding relationships with the covariates and making informed estimates of conductivity in those areas. Where those relationships are not well understood (i.e. where we see a departure in the feature space characteristics from what the model can ‘see’) the model prediction is likely to be less certain. Differences in the features and their corresponding values ‘seen’ and used in the model versus the full feature space covering the entire continent are captured in the covariate shift map. High values in the shift model can indicate higher potential uncertainty or unreliability of the model prediction. Users therefore need to be mindful when interpreting this dataset, of the uncertainties shown by the 5th-95th percentiles, and high values in the covariate shift map.
 
Datasets in this data package include:
 
1. 0_4m_conductivity_prediction_median.tif
2. 0_4m_conductivity_lower_percentile_5th.tif
3. 0_4m_conductivity_upper_percentile_95th.tif
4. 30m_conductivity_prediction_median.tif
5.30m_conductivity_lower_percentile_5th.tif
6. 30m_conductivity_upper_percentile_95th.tif
7. National_conductivity_model_shift.tif
8. Full list of referenced AEM survey datasets used to train the model (word document)
9. Map showing the distribution of training and out-of-sample sites

All the Geotiffs (1-6) are in log (10) electrical conductivity siemens per metre (S/m).
 
This work is part of Geoscience Australia’s Exploring for the Future program which provides precompetitive information to inform decision-making by government, community and industry on the sustainable development of Australia's mineral, energy and groundwater resources. By gathering, analysing and interpreting new and existing precompetitive geoscience data and knowledge, we are building a national picture of Australia’s geology and resource potential. This leads to a strong economy, resilient society and sustainable environment for the benefit of all Australians. This includes supporting Australia’s transition to net zero emissions, strong, sustainable resources and agriculture sectors, and economic opportunities and social benefits for Australia’s regional and remote communities. The Exploring for the Future program, which commenced in 2016, is an eight year, $225m investment by the Australian Government.


Reference:

Ley-Cooper, A. Y., Brodie, R.C., and Richardson, M. 2020. AusAEM: Australia’s airborne electromagnetic continental-scale acquisition program, Exploration Geophysics, 51:1, 193-202, DOI: 10.1080/08123985.2019.1694393

Wilford, J., LeyCooper, Y., Basak, S., Czarnota, K. 2022. High resolution conductivity mapping using regional AEM survey and machine learning. Geoscience Australia, Canberra. https://dx.doi.org/10.26186/146380

Lineage

Maintenance and Update Frequency: asNeeded
Statement:
The conductivity grids are derived from a covariate machine learning approach that establishers predictive correlations between the AEM conductivity depth estimates with a suite of environmental and geological grids or covariates. The conductivity depth estimates are largely derived from the regional (20 km line-spaced) AusAEM survey data (Ley-Cooper et al. 2020). Conductivity-depth values were derived from a single model using Geoscience Australia's deterministic 1D smooth-30-layer layered-earth-inversion algorithm (Brodie and Richardson 2015). The covariates include terrain attributes, gamma radiometric, satellite imagery (including Barest Earth; Wilford J. and Roberts D., 2020), gravity and magnetic derivatives, geology polygons (1:1 million surface geology) and climate surfaces. The gradient boosted tree algorithm was used to generate the prediction models. Over 460,000 training points/measurements were used in the 5-Fold cross-validation and training the model. A further 28,626 points/measurements were kept aside to assess the out of sample performance (OOS; i.e. points not used in the model training and validation). Model performance was assessed using the OOS. The OOS used label clusters or groups to ensure spatial separation from sites used in the model training. The median of the models provide the conductivity prediction and the upper and lower percentiles (95th and 5th respectively) provides a measure model uncertainty. Grids show conductivity (S/m) in log 10 units. The overall modelling approach is described in Wilford et al., 2022.

Ley-Cooper A. Y. & Brodie R. C., 2020. AusAEM: Imaging the near surface from the World’s Largest Airborne Electromagnetic Survey. In: Czarnota K., et al. (eds.) Exploring for the Future: Extended Abstracts, Geoscience Australia.
http://dx.doi.org/10.11636/134528.
 
Ross C Brodie & Murray Richardson (2015) Open Source Software for 1D Airborne Electromagnetic Inversion, ASEG Extended Abstracts, 2015:1, 1-3, DOI: 10.1071/ ASEG2015ab197
 
Wilford, J., LeyCooper, Y., Basak, S., Czarnota, K. 2022. High resolution conductivity mapping using regional AEM survey and machine learning. Geoscience Australia, Canberra. https://dx.doi.org/10.26186/146380

Wilford, J. & Roberts D., 2020. Enhanced bare earth Landsat imagery for soil and lithological modelling. In: Czarnota K.,
et al. (eds.), Exploring for the Future: extended abstracts, Geoscience Australia, Canberra,
http://dx.doi.org/10.11636/134472.

Notes

Purpose
Provides a national model prediction of surface conductivity at the 0-4 m and 30 m depth interval.The product has utility in: resource exploration, environmental assessment, infrastructure planning, groundwater studies, geological/soil mapping, research and scientific studies (i.e. providing valuable data for studying Earth's surface and near-surface processes and groundwater dynamics)

Created: 01 01 2023

Issued: 03 08 2023

Data time period: 2023-06-21 to 2023-06-22

This dataset is part of a larger collection

Click to explore relationships graph

154,-9 154,-44 112,-44 112,-9 154,-9

133,-26.5

text: westlimit=112; southlimit=-44; eastlimit=154; northlimit=-9; projection=GDA94 / geographic 2D (EPSG: 4283)

Other Information
Download the 0_4m interval data package (zip) [15.59 GB]

uri : https://d28rz98at9flks.cloudfront.net/148588/148588_00_0.zip

Download the 30m interval data package (zip) [15.77 GB]

uri : https://d28rz98at9flks.cloudfront.net/148588/148588_01_0.zip

Identifiers