|
Reporting
Period:
January 2000 - December 2000
Objectives:
1. Evaluate
how PET - based irrigation affects disease incidence and crop
water use efficiency.
2. Determine
whether multi-spectral analysis systems and other remote
sensing instrumentation can differentiate between biotic and abiotic
stresses.
A. Summary
of Progress:
Irrigation
and Disease: Take-all is a root disease of wheat caused by the soil
borne fungus Gaeumannomyces graminis var. tritici. The disease is most
prevalent under center pivot irrigation in continuous wheat fields that
are heavily irrigated. Such fields are common in the 1N-reporting district
of the Texas Panhandle. Unlike many diseases, take-all typically occurs
in the same field locations year after year. Because take-all develops
in reoccurring spots and is strongly influenced by irrigation, we felt
it was an ideal disease for management using precision agriculture technologies.
A study
was established under the center pivot irrigation system at the TAES
research site in Bushland. The center pivot is equipped with LEPA nozzles
and on-off valves. In conjunction with the on-off valves, water application
can be managed by regulating pivot speed. The pivot is also equipped
with thirty-two infrared transducers (IRTs) that are capable of measuring
crop canopy temperature. In 1998, the continuous wheat wedge of the
Bushland center pivot research site was infested with the take-all fungus.
Disease developed in infested spots in 1998 and again in 1999. In April
of 1999, take-all spots were geo-referenced for studies to be conducted
during the 2000 growing season. During 2000, differential irrigation
was applied to take-all areas and also to areas that were not infested
with the pathogen. Take-all areas that had been geo-referenced in 1999
were located and spectral readings were made weekly using the Crop Scan
hand held radiometer, and canopy temperature was recorded using the
pivot-mounted IRTs. We also contracted with a commercial remote sensing
firm, Earth Scan, to take satellite remote sensing images of our research
plots.
Unlike 1998 and 1999, take-all never developed to any significant extent
during the 2000 season. Even in plots that received 100% PET-based irrigation,
only minimal disease developed, and diseased areas were not in the same
locations as in previous years.
Therefore, no differences were detected between healthy and previously
diseased spots with any of the remote sensing techniques.
Although incidence of take-all was minimal, an unexpectedly high incidence
of Fusarium root rot developed in plots receiving 50% PET-based irrigation.
Satellite images from Earth Scan easily detected plots irrigated at
50% PET but we were unable to differentiate the difference between disease
and drought stress with the data they provided (Figure 1). We are presently
in the process of analyzing IRT data and correlating it to combine yield
monitor data.
Fig. 1.
Spectral image from the ICONOS satellite provided by Earth Scan, a commercial
remote sensing provider. Plots irrigated at 50% PET are recognized as
dark spots in the field. However, based on these images, differentiation
between drought stress and root rot was not possible.
Remote
Sensing - Differentiation between Biotic and Abiotic Stresses: If remote
sensing is to be of value for detection of plant diseases we must be
able to detect diseases before the symptoms are readily visible and
differentiate between disease and abiotic stress. Existing technologies
that are commercially available, and in use today, have limited resolution.
Thermal infrared transducers only read in the thermal infrared region,
and satellites, most handheld radiometers, and aerial imaging platforms
only use a few broad wavebands. Higher resolution data will be needed
if we are to detect spectral signatures that differentiate biotic and
abiotic stresses.
Through collaboration with Resource 21 and Boeing, we acquired a hyperspectral
radiometer capable of measuring light from 350 nm up to 2500 nm at 1
nm increments. Along with this instrument we acquired an integrating
sphere, which allows readings to be taken on a single intact leaf under
precisely controlled conditions (Figure 2). Our objective was to compare
this technology with data available from airborne imagery, hand held
radiometers, and pigment extract absorbance from a scanning spectrophotometer
with the hope of finding a unique spectral signature associated with
specific biotic stresses.
Fig. 2-
Hyperspectral radiometer (A) and an integrating sphere (B) supplied
by Resource 21 and Boeing. These imaging tools were used during the
2000 growing season to evaluate different remote sensing technologies
for their ability to differentiate between biotic and abiotic stresses.
Beet necrotic
yellow vein virus (BNYVV) causes a yellowing of sugar beet foliage that
mimics nitrogen stress. We collected samples throughout the growing
season that exhibited yellowing symptoms as well as apparently healthy
samples. Readings were taken in the field with a 9-band hand-held radiometer
and airplane and satellite images were arranged when possible. Once
the samples were in the lab they were tested for the presence of BNYVV
through ELISA. The leaves were read in the integrating sphere and then
frozen. Pigment extracts have been made on several preliminary samples
and read with a scanning spectrophotometer. When complete, the tissue
will be analyzed for nitrogen content.
The hand
held radiometer was unable to differentiate between asymptomatic plants
that tested positive for BNYVV and asymptomatic plants that tested negative
for BNYVV. Preliminary analysis of the hyperspectral data from the integrating
sphere also appeared unable to differentiate between asymptomatic plants
that tested positive for BNYVV and asymptomatic plants that tested negative
for BNYVV. However, preliminary analysis of the data from pigment extracts
showed differences between asymptomatic plants that tested positive
for BNYVV and asymptomatic plants that tested negative for BNYVV (Figure
3). This indicates that the potential exists to remotely detect disease
in the field but we are limited by our instrumentation. The sizes of
the datasets are quite large and make analysis difficult. We are currently
developing a database system, which will allow us to store and manipulate
the large datasets generated by these technologies. It will be a network
accessible, multi-user system, which will facilitate collaboration and
sharing of data between all Precision Ag faculty.
Remote
Sensing - Disease Quantification: In a separate study, we investigated
remote sensing as a means of quantitatively measuring Cercospora leaf
spot incidence and severity. The study was conducted in a Cercospora
disease nursery containing a fungicide evaluation trial; with treatments
ranging from highly effective fungicide rotations to disease check plots
with no control. An individual with years of experience in visually
rating cultivars for disease severity rated each replicated treatment
in the nursery while we used the Crop Scan handheld multi-spectral radiometer
for ratings. Preliminary results indicated that the hand-held radiometer
was a more effective means of relating disease severity to final yield
than visual disease ratings made by an experienced individual. Visual
ratings showed the highest variability at intermediate levels of disease
where individuals are less adept at estimating percentages. This work
will be repeated in 2001.
Fig. 3.
Reflectance and absorbance data from healthy and diseased Sugar beets.
Only spectrophometer readings could differentiate between Green virus-infected
tissue from green virus -free tissue.
Remote
Sensing - Definition of Management Zones - Ergot: Development of a model
for ergot risk assessment, based on historic weather records, has been
attempted before with mixed results. Although these models can be useful,
predictions are often inaccurate because of the extreme spatial and
temporal variability in environmental factors that impact disease development.
One of the basic problems with this approach is that weather data often
does not correlate well with disease incidence because weather stations
may be far removed from the fields of interest. For instance, in the
Texas Panhandle, it is not uncommon for one field to have a high incidence
of disease while fields one to five miles away might be disease free.
The weather station nearest any of these fields could be 25-50 miles
away. In such a case, it is clear that strong correlations between disease
incidence and any particular weather parameter would be a matter of
chance. Therefore, the distance of a sorghum seed production field from
the nearest weather station defines the management zone for sorghum
ergot. However, with the advent of Doppler radar, it is possible to
record meteorological events with a resolution of only a few square
kilometers, a degree of accuracy that is normally unavailable with ground
weather stations (Figure 4). National Weather Service (NWS) stations
throughout the Southern Region maintain archives of past weather events
from Doppler radar. In cooperation with the NWS in Amarillo and hybrid
sorghum seed producers, we have identified sorghum fields that were
infested with ergot during the last three years and are in the process
of relating specific weather events, as recorded by Doppler radar, with
disease. Seed companies have identified and recorded the location of
diseased fields using GPS technology. Therefore, we now have the capability
to input GPS coordinates of diseased fields into the Doppler radar program
archives and search for weather events for those specific locations
over any particular period of time. In the Texas Panhandle, widely scattered
showers frequently occur during sorghum flowering periods and these
showers may be the explanation for the variability in disease incidence.
We believe that Doppler radar technology will provide the answer to
this question.
Fig. 4.
Doppler radar images of area surrounding Amarillo, Texas. A) One hundred
mile image provides complete coverage of most of the hybrid seed sorghum
production areas in the Texas Panhandle. B) Images can be magnified
to approximately one square mile resolution, and meteorological events
for individual geo-referenced fields can be identified.
Remote
Sensing - Definition of Management Zones - Soilborne Viruses. In a second
study, we studied the distribution of BSBMV and BNYVV in sugar beet
fields in Minnesota, Colorado, and Texas. Agriculturalist working for
the various sugar companies are interested in identifying the best method
of sampling for these pathogens because plants growing in heavily infested
areas of a field can be plowed out. This selective removal of diseased
plants can result in an overall improvement in crop yield and quality.
Therefore, the size of the management zone is only limited by our ability
to describe distribution of the pathogen in the field.
Fields
infested with BNYVV, BSBMV, or both viruses were identified and grid
sampled. Grid size was one acre, for the entire field, and smaller areas
were intensively sampled with grid sizes of approximately 10'. The location
of each soil sample taken from each grid was geo-referenced using a
Satloc GPS receiver. BNYVV and BSBMV was baited from each soil sample
and bait plants tested for the presence of each virus using ELISA. ELISA
data for each geo-referenced soil sample was recorded into ArcView and
used to generate geo-referenced field maps. Data was further analyzed
using the GS+ geostatistical program to determine spatial distribution
patterns of each virus. In only one field, auto correlation existed
at a separation distance of approximately 30', but beyond this distance,
there was no observable distribution pattern (Figure 5).
Data Management: Data management is a ubiquitous problem associated
with research in precision agriculture. While all of our field data
have spatial characteristics, they are frequently in different, incompatible
units. Our yield monitor data is geo-referenced with latitude and longitude,
satellite and aerial remote sensing images are referenced according
to the universal transverse mercator system (UTM), and our infrared
thermometers that are attached to the center pivot at Bushland are geo-referenced
by distance from the pivot pad, and angle of the center pivot irrigation
system. By converting all of this data to UTM, we can visualize these
different factors.
Fig. 5.
Variograms of spatial distribution of BNYVV from intensively sampled
(small grid) section of a field (A), and the whole field in an acre-sized
(large) grid (B).
For plot-based
studies, we must also be able to locate the boundaries of our plots
to segregate these factors according to treatments and replicates. With
rectangular plots this is easily accomplished with any GIS tool. However,
with the Bushland center pivot, crops are planted in a circle to maximize
irrigation efficiency and reduce runoff. Therefore, plots are not rectangular
but small sections of a circle. By using a combination of drawings from
AutoCAD along with ArcView, we have been able to generate plot boundaries
for any given plot design (Figure 6). Whether our plots are six rows
wide or twelve, we can take the appropriate AutoCAD template, overlay
a geo-referenced ArcView image of our plots and quickly retrieve the
data of interest. This technique allows us to convert rapidly from geo-referenced
data to plot based data, thereby permitting easy statistical analysis
of any particular data set.
Fig. 6.
Data from a yield monitor combine superimposed on a plot map. Each plot
is labeled with the treatment and block. A spatial join of these images
results in a table identifying each yield point from the combine with
the treatment and block it came from.
B. Education/technology
transfer:
During
the year, members of the plant pathology project gave numerous PA presentations
at field days, crop tours, commodity research meetings, and growers
meetings. We also routinely provide tours to various groups such as
scouts, schools, visiting scientists, etc. during which we give overviews
of our PA project.
C. Milestones
achieved:
We have
made significant progress over the last three years in elucidating the
effect of various PET-based irrigation levels on crop yields, water
use efficiency and disease development. Corn, sorghum, sugar beets,
and wheat have been included in the studies. We have found that with
all grain crops, irrigation at 75% PET instead of 100% PET has typically
resulted in equal or greater potential profit to growers because of
reduced input costs. Water has been conserved and energy costs reduced.
In addition, we have found that irrigation management has the potential
to reduce the incidence and severity of several plant diseases and insect
pests. In general, reduced irrigation frequency reduces incidence and
severity of many soilborne plant pathogens more than reducing the total
amount of irrigation water applied. The most significant aspect of this
research has been the demonstration that producers cannot irrigate crops
growing in pathogen-infested soils in the same manner they irrigate
crops growing in pathogen-free soils. Soilborne plant pathogens reduce
crop yield and quality and have an adverse effect on crop water use
efficiency.
D. Publications:
Michels,
Jr., G. J., G. Piccinni, C. M. Rush, and D. A. Fritts. 1999 Sensing
Greenbug (Homoptera: Aphididae) Infestations In Winter Wheat With Infrared
Transducers. Southwestern Entomology 24(4):269-279.
Piccinni, G. and C. M. Rush. 2000. Determination of Optimum Irrigation
Regime and Water Use Efficiency of Sugar Beet Grown in Pathogen Infested
Soil. Plant Dis. 84:1067-1072
Piccinni,
G., C.M. Rush, K.M. Vaughn, and Lazar, M. D. 2000. Lack of Relationship
Between Susceptibility to Common Root Rot and Drought Tolerance Among
Several Closely Related Wheat Lines. Plant Disease. 83:25-28.
Piccinni,
G., J.M. Shriver, and C.M. Rush. 2001. The relationship among seed size,
planting date and common root rot in hard red winter wheat. Plant Disease.
(accepted)
E. Precision agriculture proposals:
Identification
and differentiation of biotic and abiotic stresses using multispectral
remote sensing for application in IPM production systems. USDA-IPM $99,000.
Integrated management practices for protecting seed sorghum from ergot.
Texas Sorghum Producers Board $25,000.
Application of precision agriculture technology for managing irrigated
sorghum at several planting densities. Texas Sorghum Producers Board
$15,000. Funded through PROFIT.
Effect of irrigation on disease incidence and severity in pathogen infested
soils. BSDF $12,200.
Potential virulence of beet soil borne mosaic virus and interactions
with necrotic yellow vein virus. Southern Minnesota ND R&D Board
$15,500. Funded
Reducing losses to take-all of wheat by remote sensing and irrigation
management. Texas Wheat Producers Board $12,000. Funded
Application of precision agriculture technology for managing irrigation
and disease in drought tolerant Corn. Texas Corn Producers Board $15,000.
Funded
Identification and differentiation of biotic and abiotic stresses using
multi-spectral remote sensing for application in IPM production systems.
NASA - Pre-Proposal $100,000.
Factors impacting development of a remote sensing dependent site-specific
irrigation/ Chemigation system. Precision Ag. - TAES $194,438. Funded.
Co-PI with Jerry Michels.
Airborne multi-spectral assessment of crop health and feedlot dust potential.
NASA $50,000. Funded
Integrated Management of Stalk Rot in Sorghum by Genetic Resistance
and Precision Irrigation. Texas Grain Sorghum Producers Board. $17,000.
Funded through PROFIT.
Development of a model for risk assessment and fungicide application
for management of sorghum ergot. Texas Sorghum Board $17,000.
Yield Tracker: A yield mapping and prediction information delivery system.
IFAFS $800,606. Cooperator, funded with Bill Payne at $40,000.
F. Precision
Agriculture meetings attended/papers (posters) presented:
G. J. Michels,
G. Piccinni, C. M. Rush, and D. A. Frits. 2000. Using infrared transducers
to sense greenbug infestations in winter wheat. Proceedings of the 5th
International Conference on Precision Agriculture. Bloomington, MN.
July 16-19,2000. (in press)
Piccinni,
G., J.K. Burk, C.M. Rush, and G.J. Michels. 1999. Development of an
Automated System for Infrared Detection of Plant Stress. p. 109. In
Agronomy Abstracts. ASA, Madison, WI. ASA/CSSA/SSSA Joint 91st Annual
Meeting, Salt Lake City, UT, Oct. 31 - Nov. 4, 1999.
G. J. Michels,
G. Piccinni, C. M. Rush, and D. A. Frits. 2000. Using infrared transducers
to sense greenbug infestations in winter wheat. 5th International Conference
on Precision Agriculture. Bloomington, MN. July 16-19,2000.
Staff Writer. 1999. Satellite, plane photos aid farmers. Houston Chronicle,
November 17, 1999, 6C.
Staff Writer.
2000. Satellite images of land take root with farmers. Albuquerque Journal.
August 14, 2000 pg. 2.
Rush, C.M.
2000. Precision Agriculture. 2000 Spring Crops Field Day. AREC Bushland,
TX.
G. Other developments:
Several personnel changes in 2000 significantly impacted the plant pathology
PA project. Keden Burk, who was farm manager for the Bushland center
pivot research site, left in early spring for another job in California.
Keden had been associated with the PA project since it began and had
assumed considerable responsibility in the overall project. His departure,
just before planting time, greatly complicated all field activities
in 2000. David Jones replaced Keden in May. David had worked in the
program as Keden's helper, but didn't have the technical or farming
experience that Keden had and this impacted our research.
Giovanni
Piccinni, an assistant research scientist who had been with the plant
pathology PA project since it began, also left early in 2000 and took
a job with TAES in Uvalde. Giovanni, a plant stress physiologist, had
taken leadership of the PA irrigation studies under the Bushland center
pivot. His departure constituted a big loss to our program but we were
still able to complete the field studies he had supervised.
Additional personnel changes that impacted the plant pathology PA project
in 2000 included the hire of two assistant research scientists. Fekede
Workneh joined our group in May. He has a Ph. D in plant pathology from
UC Davis and his experience is in epidemiology and statistics. His research
activities in PA will relate to site-specific management of plant disease
as it relates to cultural and climatic variables associated with pathogen
ecology and epidemiology. Karl Steddom was hired in June and he also
has a Ph. D in plant pathology from UC Riverside. His background is
in biological control but he has interest in data base management and
remote sensing. Karl primarily will be involved in studies evaluating
various remote sensing techniques for their ability to detect and differentiate
biotic and abiotic stresses.
|