Remote Sensing of Sagebrush Canopy Nitrogen (2011)
Investigators and Affiliations
- Jessica J. Mitchell, Ph.D. Research Affiliate, Geosciences Department, Idaho State University, Boise, Idaho
Nancy F. Glenn, Ph.D., Research Associate Professor, Geosciences Department, Idaho State University, Boise, Idaho
Teki Sankey, Ph.D., Research Assistant Professor, Geosciences Department, Idaho State University, Boise, Idaho
DeWayne R. Derryberry, Assistant Professor, Department of Mathematics, Idaho State University, Pocatello, Idaho
Matthew J. Germino, Ph.D, Supervisory Research Ecologist, United States Geological Survey, Forest and Rangeland Ecosystem Science Division, Boise, Idaho
Matt Anderson, Idaho National Laboratory, Idaho Falls, Idaho
Ryan Hruska, Idaho National Laboratory, Idaho Falls, Idaho
- Idaho Space Grant Consortium
- NOAAOAR ESRL/ Physical Sciences Division Grant # NA06OAR4600124
Remotely sensing foliar nitrogen (N) in semiarid shrublands would significantly improve our limited understanding of vegetation functionality in dryland ecosystems. Whereas vegetation indices such as Modified Soil-Adjusted Vegetation Index (MSAVI) attempt to quantify vegetation abundance, estimates of foliar N across arid and semi-arid environments could help answer process-driven questions related to topics such as controls on canopy photosynthesis and the influence of N on carbon cycling behavior. Also, in systems where soil water is the primary limiting resource and influenced by changes in available N, there is opportunity for remote sensing of foliar N to augment studies related to nutrient pulse dynamics and post-fire recovery. Remote sensing of sagebrush (Artemisia spp.) N, in particular, can yield assessments of forage nutritional status across large areas.
In 2010, investigators obtained encouraging results from applying spectroscopic methods to estimate sagebrush N concentrations in the laboratory using dry leaf material and in the field using individual live shrub canopies. The project was extended to the canopy scale using hyperspectral imagery we had the opportunity to acquire on 13 August 2010 using an airborne HyMap sensor (operated byHyVista, Inc., Sydney, Australia). This imagery augmented the hyperspectral imagery that was acquired in 2009 from an unmanned aerial vehicle (UAV) platform owned and operated by the INL (Matt Anderson and Ryan Hruska). The HyMap sensor collected two overlapping flightlines (≈ 11 km2 total) approximately 2,496 m above ground level with a nominal pixel resolution of 2.1 m. Plot level sagebrush N estimations collected in the field from 09 to 10 August 2010 were then related to corresponding reflectance data in the HyMap imagery by extracting pixels located within the plot boundaries. Several different transformation procedures were applied to the reflectance spectra and partial least square (PLS) regressions were used to identify wavelengths having a strong predictive relationship with N concentration.
- Determine if sagebrush canopy signals are strong enough to support detection of canopy N from an airborne hyperspectral platform.
- Compare how two different methods for estimating whole canopy-level N affect agreement between wavelength predictors and N concentrations in sparse desert shrubland. The first method estimates phytomass by expressing cover on a mass basis using mass per unit area (LMA) measurements while the second method uses a shrub volume surrogate that combines absolute cover and height.
- Compare the relative performance of two different spectral transformation techniques: standard derivative analysis and continuum removal.
Accomplishments through 2011
Final results from the 2010 investigation of leaf and whole canopy sagebrush N using a field spectroradiometer were published in Remote Sensing Letters. Preliminary results for estimating sagebrush canopy N concentrations from remotely sensed data were presented in oral format and published as an extended abstract for the 34th International Symposium on Remote Sensing of Environment in Sydney, Australia. Final results were submitted to Remote Sensing of Environment for publication.
Our study determined that further exploration into estimating sagebrush canopy N concentrations from an airborne platform is warranted, despite remote sensing challenges inherent to open canopy systems (Table 1). Hyperspectral data transformed using standard derivative analysis were capable of quantifying sagebrush canopy N concentrations using PLS regression with an R2 value of 0.72 and an R2 predicted value of 0.42 (n = 35). Subsetting the dataset to minimize the influence of bare ground (n = 19) increased R2 to 0.95 (R2 predicted =0.56). Ground-based estimates of canopy N using leaf mass per unit area measurements (LMA) yielded consistently better model fits than ground-based estimates of canopy N using cover and height measurements. The LMA approach is likely a method that could be extended to other semiarid shrublands. Overall, the results of this study are encouraging for future landscape scale N estimates and represent an important step in addressing the confounding influence of bare ground, which we found to be a major influence on predictions of sagebrush canopy N from an airborne platform.
Table 1. PLS Regression Results for Relating Foliar N Concentrations to HyMap Spectra Extracted from Field Plots (7 m X 7 m; n = 35). Prediction error sum of squares (PRESS) and R2 predicted values are reported.
Plans for Continuation
A final paper is expected to be published in Remote Sensing of Environment and the project is complete. Additional sagebrush nitrogen data that was collected by Matt Germino in August 2010 for experimental plots at the long-term ecological monitoring site will also be related to the HyMap imagery acquired on 13 August 2010. If feasible, the additional data will be used to further test the relationship between HyMap spectra and sagebrush canopy N concentrations.
Publications, Reports, and Theses
Mitchell, J, Glenn, N.F., Sankey, T., Anderson, M.O., Hruska, R., Hyperspectral remote sensing of sagebrush canopy nitrogen, 34th International Symposium on Remote Sensing of Environment, Sydney, Australia, April 2011.
Mitchell, J., Glenn, N., Sankey, T., Derryberry, D., Anderson, M. and Hruska, R. 2012.
Spectroscopic detection of Nitrogen concentrations in sagebrush: implications for hyperspectral remote sensing, Remote Sensing Letters, 3 (4), 285-294.
Mitchell, J., Glenn, N., Sankey, T., Derryberry, D. Hyperspectral remote sensing of sagebrush canopy nitrogen, Remote Sensing of Environment, in review.