Challenges to sensor-based N-Management for Cotton

1

• Challenges to sensor-based N-Management for Cotton

E.M. Barnes1, T. Sharp2, J. Wilkerson3, Randy
Taylor2, Stacy Worley3
1Cotton Incorporated, Cary NC 2Oklahoma State
University, Okmulgee Stillwater 3University of
Tennessee, Knoxville
2
Acknowledgements

• Tom Clarke, Glenn Fitzgerald, P. Pinter
• USDA, ARS, Arid Land Research Center
• Maricopa, AZ
• Pete Waller, University of Arizona
• Paul Colaizzi, USDA, ARS Lubbock, TX
• Julio Haberland Chile
• Mike Kostrzewski - Arizona

3
Outline

• Cotton 101
• Why cotton interest in sensors is high
• The challenges of Cotton
• One proposed solution

4
Cotton 101
5
The Cotton Belt
Data from USDA, NASS
6
Cotton Nitrogen

• Perennial plant managed as an annual
• Indeterminate flowering pattern
• 50 lbs-N per lint bale (1 bale 480 lbs)
• Over-application of N
• Energy partition to vegetative vs. reproductive
  development
• Large plants prevent efficient harvest
• Growth regulators applied to control vegetative
  development

7
Why interest in sensors now?

• Cost of N
• Producers receiving In-Time images
• And now Deere imagery through Jimmy Sanders
• On-farm tests done in Alabama to use
  GreenSeekerTM to apply growth regulator (PIX)
• Cotton researchers joining in

8
Variable Rate Defoliation
Location Arkansas Delta Crop Cotton Field
Size 339.5 Acres Imagery Acquired September
7, 2004 VR Defoliation Applied September 14,
2004 Notes This prescription was applied using
a hydraulic aerial VR system. The consultant was
able to achieve a one-time defoliation on this
field, for 15.94/A in chemical.
9
Variable Rate Nitrogen Top-Dressing
Location Arkansas Delta Crop Cotton Field
Size 156.53 As Imagery Acquired July 5,
2004 VR Fertilizer Applied July 13,
2004 Notes This prescription was applied using
a variable rate equipped high clearance spreader.
Unity 16 nitrogen (N) was applied midseason,
to supplement areas in the field which had become
N deficient. Classes 1 and 2 were beyond
salvaging with the additional N, while classes 6
and 7 required no additional N.
10
Challenges
11
Wind blows Index Changes
Heliotropic New Growth
12
Sample data set

• 1999 Growing season
• AGIIS sensor (calibration panel every minute)
• Water and Nitrogen treatments

13
1999 CCCI (relative to WN)
Last N Application
Green Boll
Open Boll
Squares
14
Yield -mNDVI C ?
15
Possible solution?
16
Combining Data

• Use NDVI / Greenseeker as a biomass sensor
• Historic yield maps.

17
Concept
18
Application
19
Theoretical Example
20
Combined
21
Conclusions

• Cotton can be tricky to manage
• Efforts to apply sensors for N management are
  increasing rapidly
• Hope to learn from work here most efficient
  methods to develop cotton N management strategies

22
AgIIS (Agricultural Irrigation Imaging System)
Bands (nm) Green (555), Red (670), Edge
(720), NIR (790)IRT
23
Field during 1999 Cotton Season
24
CCCI
A
C
CCCI (C-B)/(A-B)
B
25
1999 RVI (relative to WN)
Last N Application
Green Boll
Open Boll
Squares