Skip to main content
ARS Home » Pacific West Area » Maricopa, Arizona » U.S. Arid Land Agricultural Research Center » Water Management and Conservation Research » Research » Publications at this Location » Publication #157272
Title: DIRECTED SAMPLING USING ESAP-RSSD SOFTWARE WITH IMAGERY

Author
item Fitzgerald, Glenn
item BARNES, EDWARD - COTTON INC, CARY NC
item LESCH, S - USDA-ARS RIVERSIDE CA

Submitted to: Agronomy Abstracts
Publication Type: Abstract Only
Publication Acceptance Date: 11/4/2003
Publication Date: 11/6/2003
Citation: FITZGERALD, G.J., BARNES, E.M., LESCH, S.M. 2003. DIRECTED SAMPLING USING ESAP-RSSD SOFTWARE WITH IMAGERY. AGRONOMY ABSTRACTS. CD Rom (A03-fitzgerald156899-oral).

Interpretive Summary:

Technical Abstract: Directed sampling requires a priori knowledge of a field and it's spatial characteristics to efficiently guide soil and plant sampling. Determining number and locations of sample points has been a challenge. We used the ESAP-RSSD (ECe Sampling, Assessment, and Prediction - Response Surface Sampling Design) software to achieve this goal. Originally written to convert apparent soil electrical conductivity (ECa) measurements to soil salinity (ECe), ESAP determines the number and geo-locations of electromagnetic induction measurements required to generate a statistically robust ECe map from ECa estimtes. The algorithm assumes the observed metrics (ECa) are correlated with the measured parameter (ECe). Here, we substituted ECa with aerial imagery and ECe with plant height. Images were acquired on three dates each in 2001 and 2002 over a cotton field in Central Arizona. In 2001, from the red band of an off-the-shelf digital camera, ESAP chose twelve locations at which plant height was then measured within 48 hours of a flight. These heights were regressed with the image data to produce plant height maps. Six independent locations were also sampled for plant height to test the veracity of the calibration model. Results showed that R2 values comparing predicted and observed data ranged from 0.71 - 0.89 and were statistically different than zero. The accuracy of the prediction measured using RMSE was less than 7 cm. Thus, from twelve sample locations a plant height map was produced.