Estimation of Canopy Nitrogen Concentration Across C3 and C4 Grasslands Using WorldView-2 Multispectral Data

This paper assesses the potential of multispectral data attained using WorldView-2 (WV2) satellite to estimate and map the variability in canopy nitrogen (N) concentration across C3 and C4 grasslands. The WV2 satellite image was acquired for the Cathedral Peak region of the Drakensberg Mountain range, South Africa. The random forest (RF) regression algorithm was used to develop a relationship between two-band normalized ratio indices (RIs), including the normalized difference vegetation index (NDVI), computed from the WV2 image data and N concentration. The RF-based variable importance (VIP) scores calculated using the training dataset (n = 150) showed that the RI computed involving the costal-blue (400-450 nm) and yellow (585-625 nm) band is the most important, when predicting canopy N concentration in the area. Using the validation dataset (n = 64), the RF explained 71% of the variation, with a Nash-Sutcliffe efficiency NSE = 0.68, in predicting N across the C3 grass, Festuca costata, and C4 grasses, Themeda triandra and Rendlia altera grasslands. Overall, results indicate that predictions of canopy N using new multispectral data with unique band setting, such as WV2 spectra are possible for C3 and C4 grasslands.

[1]  M. Cho,et al.  Towards red‐edge positions less sensitive to canopy biophysical parameters for leaf chlorophyll estimation using properties optique spectrales des feuilles (PROSPECT) and scattering by arbitrarily inclined leaves (SAILH) simulated data , 2008 .

[2]  P. Curran.  Remote sensing of foliar chemistry , 1989 .

[3]  R. Barbehenn,et al.  C3 grasses have higher nutritional quality than C4 grasses under ambient and elevated atmospheric CO2 , 2004 .

[4]  P. Reich,et al.  Do species and functional groups differ in acquisition and use of C, N and water under varying atmospheric CO2 and N availability regimes? : a field test with 16 grassland species , 2001 .

[5]  S. Verzakov,et al.  Estimating grassland biomass using SVM band shaving of hyperspectral data , 2007 .

[6]  Onisimo Mutanga,et al.  Linking remotely sensed forage quality estimates from WorldView-2 multispectral data with cattle distribution in a savanna landscape , 2013, Int. J. Appl. Earth Obs. Geoinformation.

[7]  R. Phillips,et al.  Estimating forage quantity and quality using aerial hyperspectral imagery for northern mixed-grass prairie , 2007 .

[8]  Onisimo Mutanga,et al.  High density biomass estimation for wetland vegetation using WorldView-2 imagery and random forest regression algorithm , 2012, Int. J. Appl. Earth Obs. Geoinformation.

[9]  Robert W. Pearcy,et al.  II. LEAF NITROGEN EFFECTS ON THE GAS EXCHANGE CHARACTERISTICS OF CHENOPODIUM , 1987 .

[10]  Vijaya Gopal Kakani,et al.  Nitrogen deficiency effects on plant growth, leaf photosynthesis, and hyperspectral reflectance properties of sorghum , 2005 .

[11]  V. J. G. Houba,et al.  A novel digestion technique for multi-element plant analyses , 1983 .

[12]  Onisimo Mutanga,et al.  Integrating remote sensing and geostatistics to estimate woody vegetation in an African savanna , 2013 .

[13]  José M. Paruelo,et al.  Grass species differentiation through canopy hyperspectral reflectance , 2009 .

[14]  J. Schjoerring,et al.  Reflectance measurement of canopy biomass and nitrogen status in wheat crops using normalized difference vegetation indices and partial least squares regression , 2003 .

[15]  J. O S E P,et al.  Climate controls on C 3 vs . C 4 productivity in North American grasslands from carbon isotope composition of soil organic matter , 2008 .

[16]  Andrew K. Skidmore,et al.  Estimation of grassland biomass and nitrogen using MERIS data , 2012, Int. J. Appl. Earth Obs. Geoinformation.

[17]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[18]  J. Hopcraft,et al.  Herbivores, resources and risks: alternating regulation along primary environmental gradients in savannas. , 2010, Trends in ecology & evolution.

[19]  Moses Azong Cho,et al.  Mapping tree species composition in South African savannas using an integrated airborne spectral and LiDAR system , 2012 .

[20]  Onisimo Mutanga,et al.  Spectral resampling based on user-defined inter-band correlation filter: C3 and C4 grass species classification , 2013, Int. J. Appl. Earth Obs. Geoinformation.

[21]  Trevor R. Hill,et al.  Description, classification and ordination of the dominant vegetation communities, Cathedral Peak, KwaZulu-Natal Drakensberg , 1996 .

[22]  J. C. Winslow,et al.  The influence of seasonal water availability on global C3 versus C4 grassland biomass and its implications for climate change research , 2003 .

[23]  F. Baret,et al.  Leaf optical properties with explicit description of its biochemical composition: Direct and inverse problems , 1996 .

[24]  Elfatih M. Abdel-Rahman,et al.  Random forest regression and spectral band selection for estimating sugarcane leaf nitrogen concentration using EO-1 Hyperion hyperspectral data , 2013 .

[25]  Larry L. Tieszen,et al.  Climate controls on C3 vs. C4 productivity in North American grasslands from carbon isotope composition of soil organic matter , 2008 .

[26]  S. Wand,et al.  Responses of wild C4 and C3 grass (Poaceae) species to elevated atmospheric CO2 concentration: a meta‐analytic test of current theories and perceptions , 1999 .

[27]  Michele Meroni,et al.  Identification of hyperspectral vegetation indices for Mediterranean pasture characterization , 2009, Int. J. Appl. Earth Obs. Geoinformation.

[28]  David L. Peterson,et al.  Scientific issues and potential remote-sensing requirements for plant biochemical content , 1992 .

[29]  M. P. Tuohy,et al.  In-field hyperspectral proximal sensing for estimating quality parameters of mixed pasture , 2011, Precision Agriculture.

[30]  O. Mutanga,et al.  Challenges and opportunities in the use of remote sensing for C3 and C4 grass species discrimination and mapping , 2012 .

[31]  Raymond V. Barbehenn,et al.  C3 grasses have higher nutritional quality than C4 grasses under ambient and elevated atmospheric CO2 , 2004 .

[32]  Susan L Ustin,et al.  Remote sensing of plant functional types. , 2010, The New phytologist.

[33]  Onisimo Mutanga,et al.  A comparison of regression tree ensembles: Predicting Sirex noctilio induced water stress in Pinus patula forests of KwaZulu-Natal, South Africa , 2010, Int. J. Appl. Earth Obs. Geoinformation.

[34]  A. Skidmore,et al.  Nitrogen detection with hyperspectral normalized ratio indices across multiple plant species , 2005 .

[35]  Jean-Michel Poggi,et al.  Variable selection using random forests , 2010, Pattern Recognit. Lett..

[36]  Riyad Ismail,et al.  Optimizing spectral resolutions for the classification of C 3 and C 4 grass species, using wavelengths of known absorption features , 2012 .

[37]  J. Peñuelas,et al.  Remote sensing of nitrogen and lignin in Mediterranean vegetation from AVIRIS data: Decomposing biochemical from structural signals , 2002 .

[38]  U. Grömping.  Dependence of Variable Importance in Random Forests on the Shape of the Regressor Space , 2009 .

[39]  Jonathan Cheung-Wai Chan,et al.  Evaluation of random forest and adaboost tree-based ensemble classification and spectral band selection for ecotope mapping using airborne hyperspectral imagery , 2008 .

[40]  H. W. Hunt,et al.  Responses of a C 3 and C 4 perennial grass to C O 2 enrichment and climate change : Comparison between model predictions and experimental data , 2003 .

[41]  Andrew K. Skidmore,et al.  Dry season mapping of savanna forage quality, using the hyperspectral Carnegie Airborne Observatory sensor , 2011 .

[42]  Lorenzo Bruzzone,et al.  The role of spectral resolution and classifier complexity in the analysis of hyperspectral images of forest areas. , 2007 .

[43]  Alexander Hapfelmeier,et al.  A new variable selection approach using Random Forests , 2013, Comput. Stat. Data Anal..

[44]  S. Niu,et al.  Different growth responses of C3 and C4 grasses to seasonal water and nitrogen regimes and competition in a pot experiment. , 2008, Journal of experimental botany.

[45]  H. W. Hunt,et al.  Responses of a C3 and C4 perennial grass to CO2 enrichment and climate change: Comparison between model predictions and experimental data , 1996 .

[46]  R. Schulze,et al.  Incoming solar radiation patterns and vegetation response: Examples from the natal drakensberg , 1977, Vegetatio.

[47]  Xu Chu,et al.  Comparison of different hyperspectral vegetation indices for canopy leaf nitrogen concentration estimation in rice , 2014, Plant and Soil.

[48]  Andrew K. Skidmore,et al.  Hyperspectral predictors for monitoring biomass production in Mediterranean mountain grasslands: Majella National Park, Italy , 2009 .

[49]  Paul M. Mather,et al.  The role of feature selection in artificial neural network applications , 2002 .