Error propagation in spectrometric functions of soil organic carbon
<p>Soil organic carbon (SOC) plays a major role concerning chemical, physical, and biological soil properties and functions. To get a better understanding of how soil management affects the SOC content, the precise monitoring of SOC on long-term field experiments (LTFEs) is needed. Visible and...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2019-09-01
|
| Series: | SOIL |
| Online Access: | https://www.soil-journal.net/5/275/2019/soil-5-275-2019.pdf |
| _version_ | 1818390940167962624 |
|---|---|
| author | M. Ellinger I. Merbach U. Werban M. Ließ |
| author_facet | M. Ellinger I. Merbach U. Werban M. Ließ |
| author_sort | M. Ellinger |
| collection | DOAJ |
| description | <p>Soil organic carbon (SOC) plays a major role concerning chemical, physical,
and biological soil properties and functions. To get a better understanding
of how soil management affects the SOC content, the precise monitoring of
SOC on long-term field experiments (LTFEs) is needed. Visible and
near-infrared (Vis–NIR) reflectance spectrometry provides an inexpensive and
fast opportunity to complement conventional SOC analysis and has often been
used to predict SOC. For this study, 100 soil samples were collected at an
LTFE in central Germany by two different sampling designs. SOC values ranged
between 1.5 % and 2.9 %. Regression models were built using partial least
square regression (PLSR). In order to build robust models, a nested repeated
5-fold group cross-validation (CV) approach was used, which comprised model
tuning and evaluation. Various aspects that influence the obtained error
measure were analysed and discussed. Four pre-processing methods were
compared in order to extract information regarding SOC from the spectra.
Finally, the best model performance which did not consider error propagation
corresponded to a mean RMSE<span class="inline-formula"><sub>MV</sub></span> of 0.12 % SOC (<span class="inline-formula"><i>R</i><sup>2</sup>=0.86</span>). This model performance was impaired by <span class="inline-formula">Δ</span>RMSE<span class="inline-formula"><sub>MV</sub>=0.04</span> % SOC while considering input data uncertainties (<span class="inline-formula">Δ<i>R</i><sup>2</sup>=0.09</span>), and by <span class="inline-formula">Δ</span>RMSE<span class="inline-formula"><sub>MV</sub>=0.12</span> % SOC
(<span class="inline-formula">Δ<i>R</i><sup>2</sup>=0.17</span>) considering an inappropriate
pre-processing. The effect of the sampling design amounted to a <span class="inline-formula">Δ</span>RMSE<span class="inline-formula"><sub>MV</sub></span> of 0.02 % SOC (<span class="inline-formula">Δ<i>R</i><sup>2</sup>=0.05</span>). Overall,
we emphasize the necessity of transparent and precise documentation of the
measurement protocol, the model building, and validation procedure in order
to assess model performance in a comprehensive way and allow for a
comparison between publications. The consideration of uncertainty
propagation is essential when applying Vis–NIR spectrometry for soil
monitoring.</p> |
| first_indexed | 2024-12-14T05:05:36Z |
| format | Article |
| id | doaj.art-2285656b834b438385917f0f57be255d |
| institution | Directory Open Access Journal |
| issn | 2199-3971 2199-398X |
| language | English |
| last_indexed | 2024-12-14T05:05:36Z |
| publishDate | 2019-09-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | SOIL |
| spelling | doaj.art-2285656b834b438385917f0f57be255d2022-12-21T23:16:07ZengCopernicus PublicationsSOIL2199-39712199-398X2019-09-01527528810.5194/soil-5-275-2019Error propagation in spectrometric functions of soil organic carbonM. Ellinger0I. Merbach1U. Werban2M. Ließ3Department of Soil System Science, Helmholtz Centre for Environmental Research – UFZ, Halle (Saale), GermanyDepartment of Community Ecology, Helmholtz Centre for Environmental Research – UFZ, Bad Lauchstädt, GermanyDepartment of Monitoring and Exploration Technologies, Helmholtz Centre for Environmental Research – UFZ, Leipzig, GermanyDepartment of Soil System Science, Helmholtz Centre for Environmental Research – UFZ, Halle (Saale), Germany<p>Soil organic carbon (SOC) plays a major role concerning chemical, physical, and biological soil properties and functions. To get a better understanding of how soil management affects the SOC content, the precise monitoring of SOC on long-term field experiments (LTFEs) is needed. Visible and near-infrared (Vis–NIR) reflectance spectrometry provides an inexpensive and fast opportunity to complement conventional SOC analysis and has often been used to predict SOC. For this study, 100 soil samples were collected at an LTFE in central Germany by two different sampling designs. SOC values ranged between 1.5 % and 2.9 %. Regression models were built using partial least square regression (PLSR). In order to build robust models, a nested repeated 5-fold group cross-validation (CV) approach was used, which comprised model tuning and evaluation. Various aspects that influence the obtained error measure were analysed and discussed. Four pre-processing methods were compared in order to extract information regarding SOC from the spectra. Finally, the best model performance which did not consider error propagation corresponded to a mean RMSE<span class="inline-formula"><sub>MV</sub></span> of 0.12 % SOC (<span class="inline-formula"><i>R</i><sup>2</sup>=0.86</span>). This model performance was impaired by <span class="inline-formula">Δ</span>RMSE<span class="inline-formula"><sub>MV</sub>=0.04</span> % SOC while considering input data uncertainties (<span class="inline-formula">Δ<i>R</i><sup>2</sup>=0.09</span>), and by <span class="inline-formula">Δ</span>RMSE<span class="inline-formula"><sub>MV</sub>=0.12</span> % SOC (<span class="inline-formula">Δ<i>R</i><sup>2</sup>=0.17</span>) considering an inappropriate pre-processing. The effect of the sampling design amounted to a <span class="inline-formula">Δ</span>RMSE<span class="inline-formula"><sub>MV</sub></span> of 0.02 % SOC (<span class="inline-formula">Δ<i>R</i><sup>2</sup>=0.05</span>). Overall, we emphasize the necessity of transparent and precise documentation of the measurement protocol, the model building, and validation procedure in order to assess model performance in a comprehensive way and allow for a comparison between publications. The consideration of uncertainty propagation is essential when applying Vis–NIR spectrometry for soil monitoring.</p>https://www.soil-journal.net/5/275/2019/soil-5-275-2019.pdf |
| spellingShingle | M. Ellinger I. Merbach U. Werban M. Ließ Error propagation in spectrometric functions of soil organic carbon SOIL |
| title | Error propagation in spectrometric functions of soil organic carbon |
| title_full | Error propagation in spectrometric functions of soil organic carbon |
| title_fullStr | Error propagation in spectrometric functions of soil organic carbon |
| title_full_unstemmed | Error propagation in spectrometric functions of soil organic carbon |
| title_short | Error propagation in spectrometric functions of soil organic carbon |
| title_sort | error propagation in spectrometric functions of soil organic carbon |
| url | https://www.soil-journal.net/5/275/2019/soil-5-275-2019.pdf |
| work_keys_str_mv | AT mellinger errorpropagationinspectrometricfunctionsofsoilorganiccarbon AT imerbach errorpropagationinspectrometricfunctionsofsoilorganiccarbon AT uwerban errorpropagationinspectrometricfunctionsofsoilorganiccarbon AT mließ errorpropagationinspectrometricfunctionsofsoilorganiccarbon |