Modeling mid-season rice nitrogen uptake using multispectral satellite data

James Brinkhoff, Brian W. Dunn, Andrew Robson, Tina Dunn, Remy L. Dehaan

Research output: Contribution to journalArticle

Abstract

Mid-season nitrogen (N) application in rice crops can maximize yield and profitability. This requires accurate and efficient methods of determining rice N uptake in order to prescribe optimal N amounts for topdressing. This study aims to determine the accuracy of using remotely sensed multispectral data from satellites to predict N uptake of rice at the panicle initiation (PI) growth stage, with a view to providing optimum variable-rate N topdressing prescriptions without needing physical sampling. Field experiments over 4 years, 4-6 N rates, 4 varieties and 2 sites were conducted, with at least 3 replicates of each plot. OneWorldView satellite image for each year was acquired, close to the date of PI. Numerous single- and multi-variable models were investigated. Among single-variable models, the square of the NDRE vegetation index was shown to be a good predictor of N uptake (R2 = 0.75, RMSE = 22.8 kg/ha for data pooled from all years and experiments). For multi-variable models, Lasso regularization was used to ensure an interpretable and compact model was chosen and to avoid over fitting. Combinations of remotely sensed reflectances and spectral indexes as well as variety, climate and management data as input variables for model training achieved R2 < 0.9 and RMSE < 15 kg/ha for the pooled data set. The ability of remotely sensed data to predict N uptake in new seasons where no physical sample data has yet been obtained was tested. A methodology to extract models that generalize well to new seasons was developed, avoiding model overfitting. Lasso regularization selected four or less input variables, and yielded R2 of better than 0.67 and RMSE better than 27.4 kg/ha over four test seasons that weren't used to train the models.

Original languageEnglish
Article number1837
Pages (from-to)1-22
Number of pages22
JournalRemote Sensing
Volume11
Issue number15
DOIs
Publication statusPublished - 06 Aug 2019

Fingerprint

satellite data
rice
nitrogen
modeling
data management
vegetation index
profitability
train
reflectance
crop
methodology
sampling
climate
experiment

Cite this

Brinkhoff, James ; Dunn, Brian W. ; Robson, Andrew ; Dunn, Tina ; Dehaan, Remy L. / Modeling mid-season rice nitrogen uptake using multispectral satellite data. In: Remote Sensing. 2019 ; Vol. 11, No. 15. pp. 1-22.
@article{94ee754d6e4a4c3eadc65d00958bb4fc,
title = "Modeling mid-season rice nitrogen uptake using multispectral satellite data",
abstract = "Mid-season nitrogen (N) application in rice crops can maximize yield and profitability. This requires accurate and efficient methods of determining rice N uptake in order to prescribe optimal N amounts for topdressing. This study aims to determine the accuracy of using remotely sensed multispectral data from satellites to predict N uptake of rice at the panicle initiation (PI) growth stage, with a view to providing optimum variable-rate N topdressing prescriptions without needing physical sampling. Field experiments over 4 years, 4-6 N rates, 4 varieties and 2 sites were conducted, with at least 3 replicates of each plot. OneWorldView satellite image for each year was acquired, close to the date of PI. Numerous single- and multi-variable models were investigated. Among single-variable models, the square of the NDRE vegetation index was shown to be a good predictor of N uptake (R2 = 0.75, RMSE = 22.8 kg/ha for data pooled from all years and experiments). For multi-variable models, Lasso regularization was used to ensure an interpretable and compact model was chosen and to avoid over fitting. Combinations of remotely sensed reflectances and spectral indexes as well as variety, climate and management data as input variables for model training achieved R2 < 0.9 and RMSE < 15 kg/ha for the pooled data set. The ability of remotely sensed data to predict N uptake in new seasons where no physical sample data has yet been obtained was tested. A methodology to extract models that generalize well to new seasons was developed, avoiding model overfitting. Lasso regularization selected four or less input variables, and yielded R2 of better than 0.67 and RMSE better than 27.4 kg/ha over four test seasons that weren't used to train the models.",
keywords = "Lasso model, Multiple variable linear regression, Multispectral imagery, Nitrogen management, Reflectance index, Remote sensing, Rice",
author = "James Brinkhoff and Dunn, {Brian W.} and Andrew Robson and Tina Dunn and Dehaan, {Remy L.}",
year = "2019",
month = "8",
day = "6",
doi = "10.3390/rs11151837",
language = "English",
volume = "11",
pages = "1--22",
journal = "Remote Sensing",
issn = "2072-4292",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "15",

}

Modeling mid-season rice nitrogen uptake using multispectral satellite data. / Brinkhoff, James; Dunn, Brian W.; Robson, Andrew; Dunn, Tina; Dehaan, Remy L.

In: Remote Sensing, Vol. 11, No. 15, 1837, 06.08.2019, p. 1-22.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Modeling mid-season rice nitrogen uptake using multispectral satellite data

AU - Brinkhoff, James

AU - Dunn, Brian W.

AU - Robson, Andrew

AU - Dunn, Tina

AU - Dehaan, Remy L.

PY - 2019/8/6

Y1 - 2019/8/6

N2 - Mid-season nitrogen (N) application in rice crops can maximize yield and profitability. This requires accurate and efficient methods of determining rice N uptake in order to prescribe optimal N amounts for topdressing. This study aims to determine the accuracy of using remotely sensed multispectral data from satellites to predict N uptake of rice at the panicle initiation (PI) growth stage, with a view to providing optimum variable-rate N topdressing prescriptions without needing physical sampling. Field experiments over 4 years, 4-6 N rates, 4 varieties and 2 sites were conducted, with at least 3 replicates of each plot. OneWorldView satellite image for each year was acquired, close to the date of PI. Numerous single- and multi-variable models were investigated. Among single-variable models, the square of the NDRE vegetation index was shown to be a good predictor of N uptake (R2 = 0.75, RMSE = 22.8 kg/ha for data pooled from all years and experiments). For multi-variable models, Lasso regularization was used to ensure an interpretable and compact model was chosen and to avoid over fitting. Combinations of remotely sensed reflectances and spectral indexes as well as variety, climate and management data as input variables for model training achieved R2 < 0.9 and RMSE < 15 kg/ha for the pooled data set. The ability of remotely sensed data to predict N uptake in new seasons where no physical sample data has yet been obtained was tested. A methodology to extract models that generalize well to new seasons was developed, avoiding model overfitting. Lasso regularization selected four or less input variables, and yielded R2 of better than 0.67 and RMSE better than 27.4 kg/ha over four test seasons that weren't used to train the models.

AB - Mid-season nitrogen (N) application in rice crops can maximize yield and profitability. This requires accurate and efficient methods of determining rice N uptake in order to prescribe optimal N amounts for topdressing. This study aims to determine the accuracy of using remotely sensed multispectral data from satellites to predict N uptake of rice at the panicle initiation (PI) growth stage, with a view to providing optimum variable-rate N topdressing prescriptions without needing physical sampling. Field experiments over 4 years, 4-6 N rates, 4 varieties and 2 sites were conducted, with at least 3 replicates of each plot. OneWorldView satellite image for each year was acquired, close to the date of PI. Numerous single- and multi-variable models were investigated. Among single-variable models, the square of the NDRE vegetation index was shown to be a good predictor of N uptake (R2 = 0.75, RMSE = 22.8 kg/ha for data pooled from all years and experiments). For multi-variable models, Lasso regularization was used to ensure an interpretable and compact model was chosen and to avoid over fitting. Combinations of remotely sensed reflectances and spectral indexes as well as variety, climate and management data as input variables for model training achieved R2 < 0.9 and RMSE < 15 kg/ha for the pooled data set. The ability of remotely sensed data to predict N uptake in new seasons where no physical sample data has yet been obtained was tested. A methodology to extract models that generalize well to new seasons was developed, avoiding model overfitting. Lasso regularization selected four or less input variables, and yielded R2 of better than 0.67 and RMSE better than 27.4 kg/ha over four test seasons that weren't used to train the models.

KW - Lasso model

KW - Multiple variable linear regression

KW - Multispectral imagery

KW - Nitrogen management

KW - Reflectance index

KW - Remote sensing

KW - Rice

UR - http://www.scopus.com/inward/record.url?scp=85070474437&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85070474437&partnerID=8YFLogxK

U2 - 10.3390/rs11151837

DO - 10.3390/rs11151837

M3 - Article

AN - SCOPUS:85070474437

VL - 11

SP - 1

EP - 22

JO - Remote Sensing

JF - Remote Sensing

SN - 2072-4292

IS - 15

M1 - 1837

ER -