Monitoring winter wheat crop traits change during crop development using Sentinel-1 backscatters
Monitoring winter wheat crop traits change during crop development using Sentinel-1 backscatters
It is important for farmers to monitor the development changes in winter wheat (Triticum aestivum L.) crop structure to assess plant response to crop management and environmental changes. Even in cloudy conditions, radar satellite imagery has the potential to provide near real-time and reliable information about crop development that surpasses optical satellite imagery. In this study, Sentinel-1A/1B backscattering parameters for co-polarization (VV), cross-polarization (VH), differences of dual polarizations (VH-VV), and combination of dual polarizations (VH+VV) were evaluated to estimate changes in shoot density (SD), green area index (GAI), aboveground dry biomass (AGDB), plant height (PH), and leaf nitrogen content (LNC) throughout the winter wheat growing season. In order to eliminate noise caused by inverse scattering, local weighted scatterplot smoothing (LOWESS) was applied to backscatter parameters during post-processing. The correlation of backscatter parameters to the in-situ winter wheat plant traits in the fields were assessed in the Republic of Ireland and the United Kingdom over two crop growing cycles. The Sentinel-1 backscatter parameters were correlated to SD, GAI, AGDB, PH, and LNC using Support Vector Regression (SVR), Random Forest Regression (RFR), and K-Nearest Neighbors Regression (KNNR). The results presented good prediction of GAI, AGDB, PH, and LNC were possible with R2 between 0.72 to 0.95 when modelled using datasets from specific growth stages indicative of morphological changes in plant and R2 between 0.69 to 0.94 when modelled using datasets from full growth stages. It was concluded that it would be possible to develop continuous monitoring of winter wheat throughout the growing season that would not be influenced by cloud cover. Subsequent work will examine the best way to use these data for optimum crop husbandry.
Goh, Bing-Bing
6c1b543b-bdc7-4d85-b7cc-d9ed2c9fc4d3
Sattari, Sheida
7c096396-3281-4fc6-aaf8-884a498be937
Bleakley, Chris J.
b2409764-baf7-412b-a4fa-f29ed5bc3284
Holden, Nicholas
d2594a04-d45b-4bd7-98c7-18041826a897
18 April 2023
Goh, Bing-Bing
6c1b543b-bdc7-4d85-b7cc-d9ed2c9fc4d3
Sattari, Sheida
7c096396-3281-4fc6-aaf8-884a498be937
Bleakley, Chris J.
b2409764-baf7-412b-a4fa-f29ed5bc3284
Holden, Nicholas
d2594a04-d45b-4bd7-98c7-18041826a897
[Unknown type: UNSPECIFIED]
Abstract
It is important for farmers to monitor the development changes in winter wheat (Triticum aestivum L.) crop structure to assess plant response to crop management and environmental changes. Even in cloudy conditions, radar satellite imagery has the potential to provide near real-time and reliable information about crop development that surpasses optical satellite imagery. In this study, Sentinel-1A/1B backscattering parameters for co-polarization (VV), cross-polarization (VH), differences of dual polarizations (VH-VV), and combination of dual polarizations (VH+VV) were evaluated to estimate changes in shoot density (SD), green area index (GAI), aboveground dry biomass (AGDB), plant height (PH), and leaf nitrogen content (LNC) throughout the winter wheat growing season. In order to eliminate noise caused by inverse scattering, local weighted scatterplot smoothing (LOWESS) was applied to backscatter parameters during post-processing. The correlation of backscatter parameters to the in-situ winter wheat plant traits in the fields were assessed in the Republic of Ireland and the United Kingdom over two crop growing cycles. The Sentinel-1 backscatter parameters were correlated to SD, GAI, AGDB, PH, and LNC using Support Vector Regression (SVR), Random Forest Regression (RFR), and K-Nearest Neighbors Regression (KNNR). The results presented good prediction of GAI, AGDB, PH, and LNC were possible with R2 between 0.72 to 0.95 when modelled using datasets from specific growth stages indicative of morphological changes in plant and R2 between 0.69 to 0.94 when modelled using datasets from full growth stages. It was concluded that it would be possible to develop continuous monitoring of winter wheat throughout the growing season that would not be influenced by cloud cover. Subsequent work will examine the best way to use these data for optimum crop husbandry.
Text
SSRN-id4422047
- Author's Original
More information
Published date: 18 April 2023
Identifiers
Local EPrints ID: 485108
URI: http://eprints.soton.ac.uk/id/eprint/485108
PURE UUID: ea33a98c-b0bb-4653-99b9-494a5e8c3819
Catalogue record
Date deposited: 29 Nov 2023 17:55
Last modified: 18 Mar 2024 04:12
Export record
Altmetrics
Contributors
Author:
Bing-Bing Goh
Author:
Sheida Sattari
Author:
Chris J. Bleakley
Author:
Nicholas Holden
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics