Quantifying Stripe Rust Reaction in Wheat Using Remote Sensing Based Hand-held NDVI Sensor

  • Published on
    28-Aug-2014

  • View
    674

  • Download
    2

Embed Size (px)

DESCRIPTION

Apoorva Arora

Transcript

<ul><li> Quantifying stripe rust reaction in wheat using remote sensing based hand held NDVI sensor Apoorva Arora Directorate of Wheat Research, Karnal </li> <li> Normalized dierence vegeta2on index (NDVI) Func2on of incident and reected light NDVI= NIR Red , NIR + Red NIR 750-1300 nm Red 600-700 nm Foliar pigments dominate reectance measurements 0&lt; NDVI&lt; 1 </li> <li> Stripe Rust and NDVI Breakdown of foliar pigments Foliar physiological ac2vity decreases Less reactance of infrared by healthy vegeta2on and vice versa Decrease in value of NDVI NASA Earth Observatory (Illustra2on by Robert Simmon) </li> <li> Why NDVI Most sensors provide measurements in NIR &amp; Red por2on of spectrum Addi2ve gene governed rust resistance geZng more a[en2on than ever before Minor varia2ons can be captured by quan2ta2ve varia2on </li> <li> High Throughput Screening Gap Phenotyping Genotyping Phenotyping in eld condi1ons?? </li> <li> Rust Scoring Methodology </li> <li> Satellite based Remote sensing: Limita2ons Atmospheric condi2ons Satellite geometry &amp; calibera2on Soil backgrounds Crop canopy Angle of solar radia2on incidence Small plots cant be used for measurements </li> <li> Objec2ve . Quantifying stripe rust reaction in wheat using remote sensing based hand held NDVI sensor </li> <li> Research Methodology </li> <li> Study Material 322 genotypes Pedigree analysis Diversity analysis 120 genotypes Released varie1es, Local land races,Elite wheat genotypes Represen1ng each group </li> <li> Field Experiment LaZce Design </li> <li> Plan2ng Method </li> <li> Epiphyto2c Condi2ons Epiphyto2c condi2ons created by plan2ng suscep2ble check on either sides of plot Inoculated with Yr27 virulent race 78S84 of Puccinia striiformis </li> <li> Equipment Recorded data using handheld ac2ve op2cal GreenSeeker sensor (Trimble Industeries, Inc.) NDVI computed from reectance measurements in red (~660nm)&amp; near infrared (around 780nm)por2on of spectrum Display value in range of 0.00 to 0.99 </li> <li> RESULTS </li> <li> AUDPC wise distribu2on of genotypes AUDPC computed varied from 0 to 2077 Equal number of genotypes in all categories except in AUDPC range of 1-100 </li> <li> NDVI Recorded when crop showed symptoms of maximum infec2on Value of NDVI varied from 0.46 to 0.69 Mean values across dierent AUDPC range: 0.58 to 0.69 Value reduced with increase in incidence of disease </li> <li> NDVI vs AUDPC Regression equa2on for NDVI using AUDPC : NDVI=0.663+-6.165E-5(AUDPC)t Signicant correla2on depicted by r2 value of 0.63 NDVI AUDPC </li> <li> Eect of Spot Blotch Spectral quality of reected light from leaves manifested in leaf color NDVI values also got aected by the presence of spot blotch Quan2ed value of NDVI due to blotch alone was added (if required) </li> <li> Mean NDVI values amer correc2on: 0.64 to 0.76 Regression equa2on for Corrected NDVI vs AUDPC: NDVI=0.738 +-7.061E-5(AUDPC)t Correla2on (r2 ) value improved to 0.69 NDVI AUDPC </li> <li> Plant physiological factors AUDPC Correla2on coecient showed signicant values Correla2on value improved as categories shimed from predominantly resistant to suscep2ble types AUDPC Range Coe. of determina2on (r2) Correla2on coecient 0-200 0.20 -0.45 &gt;200 0.72 -0.85 </li> <li> Plant Height Plant Height (in cm) No. of genotypes Coe. of determina1on (r2) Correla1on coecient 65-74 3 0.76 -0.87 75-84 15 0.62 -0.79 85-94 50 0.62 -0.79 95-104 35 0.73 -0.85 105-114 14 0.57 -0.76 115-124 2 1.00 -1 Value of correla2on coecient increased as taller types were more suscep2ble Range of coecient : 0.76 to 1.00 </li> <li> Similar value of correla2on coecient obtained No signicant dierence between categories was no2ced Waxiness and Early Growth Habit </li> <li> Conclusion With increasing a[en2on towards quan2ta2ve rust resistance studies, innova2ve tools &amp; techniques are needed NDVI sensor technique provides mean value of several images captured from the plot as against single frame observa2on by human eye High correla2on value indicates suitability of the instrument as an useful tool for accurate rust data recording Accuracy of this method improves when catagoriza2on of genotypes accrued to predominance of susep2bility </li> <li> Inspiration behind work </li> <li> Dr. Indu Sharma Project Director Directorate of Wheat Research Karnal Dr. M. S. Saharan Principal Scien1st Dr. R.K. Sharma Principal Scien1st Dr. K. Venkatesh Scien1st Davender Sharma Sr. Research Fellow </li> <li> Thank You </li> </ul>