Conversion & Calibration:: Cassie Knierim

Conversion & Calibration:

Cassie Knierim Student Airborne Research

Overview vBac kg rou nd In for mati on vIm po rtance vObjec ti ve s

vData C ol lecti on vMAST ER and i n the fi eld

vAnaly si s vCon ver ting fr om radi ance to temp er at ur es vCali br at ion vMo de ls to ap pr oxi mate cano py temp er at ur e

vConcl usion

Importance v Stu die s sh ow th e le af tem pe ratu res of healthy plants will be a few ⁰C less th an a ir t em per ature . v In w ater -st re ssed agr icult ur al regi ons , li ke Ke rn Coun ty, CA, eff ici ent and pr eci se irriga tion pl ans are re qu ired. v Nu mer ous indi ces ex ist for de ter minin g rela tiv e p la nt hea lth vmany requi re accur at e cano py

Local Thermal Measurements

Remote Thermal Measurements

Advantages:

Advantages:

6. Better accuracy (less atmospheric effects, etc.) 7. Better spatial resolution

5. Increased field of view • Better for irrigation plans of large field 7. Free satellite images available daily

Disadvantages: 12.Small field of view • Inefficient for irrigation planning in larger fields 14.Thermal guns are expensive

Disadvantages 11.Poor spatial resolution 12.More calibration required for accurate results

Objective vCon vert a nd ca li br ate th e rem otely sen sed da ta f rom th e M ASTER i nto accu ra te can op y tempe ratu res for th e a lmon d f ield for u se i n stu dy ing crop h ealth .

Data Collection In the Field

St udy si te: Almo nd Fi eld s near Lost H ills, Califor nia Fi eld Day: 7/2 2/0 9

Data Collection

C AL1

In the Field

Cali br at ion Si tes : CA L1, CA L2 = Bare Soi l CA L3, CA L4 = W ate r

C AL2

Mea su red te mper atur e in a gr id p att ern (~ 10m 2 /g rid bo x)

1. Re co rd time and lo cat ion. T1

T4

T7

T2

T5

T8

T3

T6

T9

C AL3

3. Re co rd scale fact or us in g b lackbo dy cali brat io n t ar get . 5. Measur e t empe rat ur e o f each gridbo x (d id no t us e cent er g rid fo r wat er si tes) • boxe s con tain ing

C AL4

Data Collection In the Field

Measured the temperature of the canopy using a handheld thermal gun and a bucket truck. Measurement Pattern (example): Location: 226W (truck is on the West side of Tree 226) Column Column Column Column Column

1: 2: 3: 4: 5:

1 1 1 1 1

2 2 2 2 2

3 3 3 3 3

Full Canopy Mixed Canopy Soil/Shade Mixed Canopy Full Canopy

4 4 4 4 4

5 5 5 22 5 6 5

Analysis General Model L0 = radiance emitted

by the atmosphere MASTER

L

τ(1-ε)

g

Lg

Lg = radiance that L 0

τεLB

reaches surface from sun LBB = radiance of

blackbody at same temperature as surface ε = emissivity of surface

Analysis Getting radiance from surface

Atmospheric Correction

MASTER

L

τ(1-ε)

g

Lg

(MODTRAN 4)

L 0

τεLB

Lg from MODTRAN 4 ε estimated (others’

Analysis Getting temperatures

T = temperature in Kelvin, λ = wavelength of the band center in meters h = Planck’s constant (6.626068×10-34m2kg/s) c = speed of light (299792458 m/s) k = Boltzmann’s constant (1.3806503×10‐23 m2 kg/s2 /K) LBB = radiance in W/m2/sr/m

Analysis Table 1. Raw data collected in the field on 7/22/2009 using a handheld thermal gun. Time: Universal Coordinated Time (UTC). TIR: scale factor for calibration of thermal gun temperatures. Green represents algae.

Analysis Table 3. Scale factor applied to ⁰C temperatures and then converted to Kelvin (+273.15). Algae grids removed. Averages and standard deviation for each calibration site and time calculated.

Analysis

Plot of average calibration temperatures for each site at two different times. Linear trendline plotted to approximate change in temperature over time (assumed linear change).

Analysis Calculating Temperatures from MASTER data

Note: Peak em it ted wa vel engt h for Eart h falls in to Band 42.

Analysis Calculating Temperatures from MASTER data

Analysis Rough Validation

Temperature Image of Research Orchard

From Field Measurements

Calibrated Temperatures appear to be reasonable based on the field measurements (pixel temperature is a weighted average of soil and canopy temperatures,

Analysis

Results: Calibrated Temperatures Calibrated Temperatures of Almond Orchards

Difference Against Uncalibrated Temperatures with assumed ε

Analysis Results -3 0

0 10

Ai r Te mp. – Pi xel Tem p. Ai r Te mp. at Tim e of Fl igh t: 304. 82 K (fr om Bel ri dge Sta ti on )

Limitations vField data coll ected on 7/2 2, MASTE R da ta on 7 /24 vLi mi ted ca nop y t emperatur e mea su remen ts fo r vali dati on pur poses vPossi ble error in atm osp her ic correcti on vCali br ati on li ne fr om only 4 poi nts

Conclusion vCa li br ation vCa li br ation wi th fi el d measur eme nts is requ ire d for accurate canopy te mper atu res vMa ny fa ctors aff ec t th e re su lt s inc ludi ng: a tm osph er ic effe cts, pe rc ent c anop y c ove r ver su s per cent sha ded and pe rc ent su nny soil , em issi vi tie s, geom et ry of th e c rops, su n a ngle , sen sor loc ati on/ angl e, et c. vResul ts vThr ee of t he fou r fie ld s had pi xel te mper atu res less tha n a ir

Further Research vImpr ove m odels fo r at mo sph er e and char act eri st ics o f sur face vStud ied a mo del de si gned fo r orang e grov es that rep ort ed si gni fi cant accur acy , but lack of values fo r inp ut p ar amet ers and po ss ib le inap pli cab il it y cause d it to fai l vApp ly “unmi xi ng ” result s (app roxi mat e pe rcent ag e of co mpo nen ts in e ach pix el ) vAcqu ir e mo re fi eld dat a fo r imp rov ed cali br at ion and mo del vali dati on

Acknowledgements & References Th ank s to Sha wn & Ni ck, and th e re st of the ET grou p for a ll the ir he lp and adv ice! Th ank s to ev er yon e who ma de SARP possible ! MAST ER Wo rk sho p Ha ndboo k, 200 9, Cent ro Nac iona l de Al ta T ecnolo gía A Phy sic al M odel fo r Inter pre ting t he La nd S urfa ce Temper atu re Obta ine d by Remo te Senso rs ov er In complete C ano pies, Vice nt e Ca sell es, Jo se A . So bri no , and C esa r C oll

Questions?