Winter cropping control

Owned by Maiwenn Herpe
Last updated: May 31, 2021 by Frederika Mourot
4 min read

Satellite(s)

Sentinel-2, Sentinel-1, Landsat, Planet Dove, Skysat, WorldView.

Monitoring element

Land surface reflectance, SAR backscatter.

Satellite(s)

Sentinel-2, Sentinel-1, Landsat, Planet Dove, Skysat, WorldView.

Monitoring element

Land surface reflectance, SAR backscatter.

Description technique

Winter cropping control is a specific application that uses a similar range of techniques as described in the land cover change detection section, specifically using a time series of NDVI or similar vegetation index to map areas of significant change within active farmland during the winter months.
The challenge for detecting winter cropping tillage or overgrazing, comes from persistent cloud cover. This limits the availability of 'clean' time series of optical satellite data, which entails long term time series analysis and winter agriculture monitoring potential (Baldi et al, 2008, Denize et al, 2019).
Cloud and cloud shadow masking has been an area of significant and ongoing development for open access optical sensors, particularly Sentinel-2 and Landsat (e.g. Baetens et al, 2019). This has allowed vegetation indices to be applied through winter using the available cloud free pixels, allowing detection of winter grazing via comparison of subsequent image vegetation index values, or mapping areas of high variance through winter etc.

Time series of SAR images have also been used to map bare soil relative to other agricultural vegetation classes over winter, with differences in backscatter across available polarisations (Ho Tong Minh et al, 2018).

Manual or automated delineation from aerial imagery should be possible, but is also weather dependent. While it is possible to fly below the cloud layer, this can increase costs due to additional flight times necessitated by lower altitudes.

Accuracy / Resolution

Variable spatial and temporal resolution according to sensors.

Case study

Ho Tong Minh et al. (2018) demonstrate the use of Sentinel-1 SAR time series to classify agricultural vegetation quality into five classes, from high to bare soil, using a recurrent neural network (RNN) model. The results suggest such a technique would be readily applicable to winter cropping control.

Indufor has developed a proof of concept for detecting winter over-grazing using a Sentinel-2 time series in the Waikato Region to satisfy National Environmental Standards monitoring requirements. The technique utilised a NDVI signal to monitor for significant changes in pasture cover within areas defined by the NZ Land Use Map dataset as productive land:

https://www.indufor.co.nz/blog/post/eo-tools-support-catchment-scale-environmental-monitoring

Benefits

  • Provides a consistent synoptic overview at significantly reduced cost when compared to field monitoring while avoiding the need to set up site visits with landowners.

  • The use of high revisit data such and Sentinel-1 and Sentinel-2 (as well as commercial data such as Planet) allows for much higher temporal granularity and regularity than what may be achieved by any other means.

  • Time series change detection can be combined with a range of other geospatial data to create additional metrics, such as terrain models (include slope classes in analysis), stream layers (include distance to streams from areas of detected winter cropping) etc.

Limitations

  • Despite best efforts with cloud and cloud shadow masking, there may be a fundamental lack of cloud free pixels over the target environment, which will result in monitoring gaps. While these may be filled with averaged or otherwise interpolated values, this still hinders change detection techniques (e.g. Baldi et al, 2008).

  • Topographic shadows present during winter due to low sun angle can represent a significant barrier to automated detection of land cover change within topographically complex areas.

Applicability for Northland

Yes, likely.

Techniques applying optical data will be limited in coverage and temporal granularity by the persistent cloud cover in the region, particularly during the winter months. Mature cloud-masking techniques are directly available for open access multispectral data (e.g. Landsat and Sentinel-2). When using commercial data, care must be taken to ensure that there is sufficiently cloud free imagery available, as cloud masking is not as mature and ordering a large volume of imagery to ensure complete cloud free coverage between multiple observations can become cost prohibitive.

Recent developments with cloud masking and online processing platforms have reduced the technical barriers for developing time series monitoring applications.

Publication references

Tian, H., Huang, N., Niu, Z., Qin, Y., Pei, J. and Wang, J., 2019. Mapping winter crops in China with multi-source satellite imagery and phenology-based algorithm. Remote sensing, 11(7), p.820.
https://www.mdpi.com/2072-4292/11/7/820

Baldi, G., Nosetto, M.D., Aragón, R., Aversa, F., Paruelo, J.M. and Jobbágy, E.G., 2008. Long-term satellite NDVI data sets: evaluating their ability to detect ecosystem functional changes in South America. Sensors, 8(9), pp.5397-5425.
https://pubmed.ncbi.nlm.nih.gov/27873821/

Denize, J., Hubert-Moy, L., Betbeder, J., Corgne, S., Baudry, J. and Pottier, E., 2019. Evaluation of using sentinel-1 and-2 time-series to identify winter land use in agricultural landscapes. Remote Sensing, 11(1), p.37.
https://www.mdpi.com/2072-4292/11/1/37

Minh, D.H.T., Ienco, D., Gaetano, R., Lalande, N., Ndikumana, E., Osman, F. and Maurel, P., 2018. Deep recurrent neural networks for winter vegetation quality mapping via multitemporal SAR Sentinel-1. IEEE Geoscience and Remote Sensing Letters, 15(3), pp.464-468.
https://arxiv.org/abs/1708.03694

Baetens, L., Desjardins, C. and Hagolle, O., 2019. Validation of Copernicus Sentinel-2 cloud masks obtained from MAJA, Sen2Cor, and FMask processors using reference cloud masks generated with a supervised active learning procedure. Remote Sensing, 11(4), p.433.

https://www.mdpi.com/2072-4292/11/4/433