Source: UNIVERSITY OF CALIFORNIA, BERKELEY submitted to NRP
GROUND-PENETRATING RADAR: DEVELOPMENT OF A VINEYARD MANAGEMENT TOOL
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
COMPLETE
Funding Source
NRI COMPETITIVE GRANT
Reporting Frequency
Annual
Accession No.
0186897
Grant No.
2001-35102-09866
Cumulative Award Amt.
$250,000.00
Proposal No.
2000-00958
Multistate No.
(N/A)
Project Start Date
Nov 15, 2000
Project End Date
Nov 14, 2005
Grant Year
2001
Program Code
[(N/A)]- (N/A)
Recipient Organization
UNIVERSITY OF CALIFORNIA, BERKELEY
(N/A)
BERKELEY,CA 94720
Performing Department
CIVIL AND ENVIRONMENTAL ENGINEERING
Non Technical Summary
There are currently no techniques available to yield information about soil heterogeneities and water content at both the resolution and spatial coverage needed to assist crop development and management, and in particular, vineyard management. Prompted by successful results of a controlled Ground Penetrating Radar (GPR) pilot study, we propose to test the applicability GPR methods to interpret soil variabilities and to estimate water content under real field conditions at a study site within the Robert Mondavi vineyards in Napa County, California. GPR is a surface-based geophysical tool that uses high frequency electromagnetic energy to probe the subsurface. Variations in electromagnetic properties, inferred from changes in the recorded signal, can be indicative of soil texture or water content variations in both the lateral direction as well as with depth. GPR methods perform optimally over the rooting depth and moisture content ranges common for grapevines, and has the capability to provide the necessary soil and water content information in multi-dimensions, with high resolution, and in a non-invasive manner. Detailed soil variability and water content at a vineyard site could assist in estimation of plantable acreage, in the design of vineyard layout and in the design of a more efficient irrigation scheme. Our current research involves investigation of GPR acquisition, inversion and interpretation techniques under different field conditions to determine the viability of this method as a reliable and efficient water-content field tool. A natural extension of the
Animal Health Component
20%
Research Effort Categories
Basic
80%
Applied
20%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
1110210203020%
1111131205020%
1020120310020%
4040210202040%
Knowledge Area
102 - Soil, Plant, Water, Nutrient Relationships; 111 - Conservation and Efficient Use of Water; 404 - Instrumentation and Control Systems;

Subject Of Investigation
0120 - Land; 1131 - Wine grapes; 0210 - Water resources;

Field Of Science
2020 - Engineering; 2030 - Geology; 2050 - Hydrology; 3100 - Management;
Goals / Objectives
The objective of our research is to develop surface Ground Penetrating Radar (GPR) methods for use as a field tool to provide information about soil heterogeneity and moisture content. GPR is a geophysical tool that uses high frequency electromagnetic signals to probe the subsurface. By investigating changes in travel time and amplitude of the recorded signals in space and time, we will investigate methods to estimate soil texture and changes in soil water content. The focus of our research involves investigating different ways of acquiring and interpreting GPR measurements in terms of these soil parameters, as well as integration of the spatially dense geophysical information with sparse, one-dimensional borehole or soil sample measurements. We have already tested the GPR concept under ideal conditions as a constructed test site and have found that GPR methods have the capability to provide high-resolution, multi-dimensional information about the soil parameters in a non-invasive manner. We are now prepared to teset the concept under field conditions involving heterogeneous media and different states of saturation. We will perform our work at the Robert Mondavi Vineyards in Napa County, California, where within-block vegetation and fruit quality variabilities exist and where other researchers are investigating the utility of plant-and remote sensing information as aids to vineyard management. We intend the soil heterogeneity and moisture content information obtained from GPR to be useful for several aspects of vineyard management including: assessment of plantable acreage, layout of vineyard blocks and design of irrigation and agrochemical application schemes. After we investigate the technical aspects of using GPR to obtain the necessary soil information we will assess the practicality of using the GPR methods as a field tool. This assessment will include testing different GPR acquisition platforms that could facilitate GPR data acquisition, and testing the concept of using information from selected control sites within a geostatistical framework to infer soil parameters spatially over larger areas. Collecting our soil-based measurements in tandem with other researcher's plant- and airborne-based measurements at the same site also leaves us well poised to assess and compare in future studies the utility of the different types of measurements for precision vineyard management.
Project Methods
The proposed development of GPR methods into a field-based soil water content and heterogeneity tool will be carried out through the five research tasks described below. Taks 1: Acquisition of Initial Data Set At the Mondavi site, we will test several different GPR acquisition geometries and parameters to determine which settings are optimal for imaging the soil heterogeneities and soil water content information at the spatial resolution necessary to aid vineyard operations. Once suitable acquisition geometries and parameters are obtained, we will collect a grid of radar profiles, install neutron probe access tubes, and collect soil samples for textural analysis. Task 2: Initial Data Set Analysis Calibration. Prior to comparison with radar data, the neutron probe measurements must be calibrated to water content and the GPR reflectors must be interpreted in terms of interfaces between soil layers having different textures. These reflectors, once identified, will be mapped to provide three-dimensional cube of soil layer information at the study site. Petrophysical Relationships. To convert geophysical attributes into soil water content, reliable petrophysical relationships are needed. We propose to investigate this issue from both a laboratory perspective and from a field data-based approach. Interpretation Techniques. After the petrophysical relationships are developed, we will use three different techniques to extract radar and velocity information from the radar data that will in turn be used to estimate water content. Development of these techniques represents the crux of our investigation. The first method involves using changes in reflected energy velocity obtained from GPR data collected at the same location but at different times to infer changes in water content as a function of distance and depth. The second method will involve investigating variations in ground wave velocity to infer an effective moisture content of the near surface. Finally, we will investigate the relationship between recorded signal amplitude, magnitude and polarity and moisture content variations. Once we have determined which of the three GPR interpretation method(s) is the most reliable and efficient for use in the vineyard, we will convert the GPR survey into estimates of water content. Taks 3: Delineation of Control Sites using Geostatistics. We shall perform geostatistical analysis of the water content estimates and soil properties to determine models of spatial variability. These models and the remote sensing data will be jointly used to design the quantity and location of control sites where, in practice, subsequent GPR and neutron probe data could be collected in a time-lapse fashion. Use of a limited number of control sites reduces the amount of data needed to describe the variations for a given site, and thus reduces cost and effort associated with this method. Task 4: Time-Lapse Data Acquisition and Interpretation We will collect radar data and neutron probe data three times during the year under different moisture conditions at the study site.

Progress 10/01/02 to 09/30/03

Outputs
Our USDA study involved investigating the utility of a surface geophysical technique called Ground Penetrating Radar (GPR) for providing dense estimates of soil water content in a non-invasive manner. The GPR system is about the size of a vacuum cleaner, and can be manually pulled through the crop rows, or can be hooked to the back of farm machinery. The GPR transmitter sends an electromagnetic wave into the subsurface, which travels through the soils and is subsequently recorded by a receiver. The velocity of the GPR waves traveling in the near subsurface soils is proportional to water content. The objective of the project was to ascertain the accuracy and resolution of GPR-obtained estimates of soil water content. Field experiments were collected during 2000-2002 at the Robert Mondavi Winery in Napa County, and during 2002-2003 at the Dehlinger vineyards in Sonoma County, California. We chose to test the approach within vineyards, because it is a significant cash crop, and the crop quality is tightly linked to water availability. The largest agricultural market in the USA is in California, and wine is the state's most valuable finished agricultural product, with an industry valued at $1.7 billion (in 1997). Most of California's agricultural production would not be possible without irrigation. However, California uses the largest volume of water of any state in the nation, and is on the verge of a major water shortage. As vineyards consume more rural acreage, competition for water resources is increasing, which has increased the pressure on California vintners to use water more efficiently. In addition to the impact of irrigation on water resources, the volume and timing of irrigation has a great impact on winegrape quality and yield. Current techniques of measuring soil moisture typically involve sampling the soil at a few spot locations within a vineyard. Not only are such techniques costly and invasive, they may not always create an accurate representation of the vineyard soil moisture distribution since the soil at one location may be quite different from the soil a few meters away. Through testing conducted within the California vineyards, we have shown that analysis of both GPR ground and reflected wave velocities can provide accurate estimates of soil water content with a spatial resolution that is unparalleled by conventional measurement techniques. The accuracy of the 100 MHz groundwave estimates of water content obtained at the Mondavi Winery were on the order of 1% compared to conventional and invasive 'point' measurements of water content, and the accuracy of soil moisture estimates obtained from 100 MHz GPR reflection data was on the order of 3%. The groundwave samples the near-soil layer, while the reflection mode samples an effective zone between the ground surface and the depth of reflecting horizons, such as soil layer interfaces. We estimated that the 900 MHz groundwave data sampled the top 15cm of the soil surface, and the reflecting horizon that we used to map the effective soil layer water content varied in depth from 0.9-1.6m below the ground surface. We found that there was a spatial pattern of soil moisture content, which was controlled in the very near surface by soil texture and in the deeper soil layers by a combination of seasonal precipitation, geologic structure, and topography. Although both the groundwave and reflection approach offer reasonable accuracy for precision agriculture, the GPR groundwave approach offers the potential for quickly assessing the near surface soil moisture in a straightforward manner, while the reflection approach requires more calibration and thus more effort. As such, the GPR groundwave approach is better suited for a field moisture sensing tool. The use of both types of arrivals (as well as multiple frequency GPR data) can be used to create 3D soil water content estimates. Although the technique was developed and tested at California vineyards, it has wide applicability to any area where it is beneficial to understand soil water distribution. For example, the GPR approach could be used to guide precision irrigation of other agricultural crops (such as orchards), as well as for input to ecosystems studies and soil-vegetation-atmosphere transfer models.

Impacts
By providing detailed information about soil moisture, farmers can refine their irrigation strategies to produce an improved agricultural product while saving water resources. For example, the technology can be used to help vineyard managers decide how to optimally develop new vineyards; vineyards could be planted within areas where the soil is spatially uniform, and managers could further refine the practice of matching grape variety to soil conditions. There has been much interest within both the scientific and the agricultural community about the results of this USDA study. The results and significance of the study to precision viticulture was nationally publicized via a Next at CNN report, which can be viewed at: . The study results have also been reported by several journals, such as the Wine Spectator, Wine Business, California Agriculture, and the Economist, and it has also recently served as a NRI Research highlight. The ability to obtain dense estimates of soil moisture holds potential for improving both precision viticulture and ecosystem investigations. The series of well-received publications testifies to the significant impact of this USDA funded research. Beyond the scientific progress are long-term societal contributions associated with this USDA award, such as: (1) One of the four graduate students/researchers involved in the project was Asian and two were women, (2) in conjunction with the UC Berkeley College of Engineering undergraduate research opportunities Program, the USDA-funded research provided projects for four underrepresented undergraduate students from UCB Civil Engineering, and (3) the USDA support was significant in training a new group of students in the multidisciplinary field of hydrogeophysics.

Publications

  • Grote, K., Hubbard, S., and Rubin, Y., 2001, Soil Water Content Spatial Correlation Estimation using GPR, Eos. Trans. AGU 82(47), Fall Meet. Suppl., Abstract H31C-0247.
  • Kowalsky, M.B., and Y. Rubin, 2001, Improved characterization of the Vadose zone with time-lapsed ground-penetrating radar, EOS. Trans. AGU 82(47) Fall Meet. Suppl.
  • Hubbard, S. Grote, K. and Rubin, Y., 2002, Estimation of near-subsurface water content using high frequency GPR ground wave data, Leading Edge of Exploration, Society of Exploration Geophysics, v21(6), 552-559.
  • Kowalsky, M.B., Rubin, Y., and Finsterele, S.A., 2002, Inversion of Hydrogeological and Time-lapsed GPR Data for Flow Parameters, AGU, 83(47), Fall Meeting Suppl., Abstract H61A-0750.
  • Zhangshuan Hou, Yoram Rubin, and Susan Hubbard, 2002, Numerical Simulation of Soil Water Content in Vadose Zone Using Constraints Provided by Geophysical Measurements, Eos. Trans. AGU, 83(47), Fall Meeting Suppl., Abstract H61A-0746.
  • Kowalsky, M.B., P. Dietrich, G. Teutsch, and Y. Rubin, 2001, Forward modeling of ground-penetrating radar data using digitized outcrop images and multiple scenarios of water saturation, Water Resources Research, Vol. 37, No. 6 , p. 1615.
  • Kowalsky, M.B., and Y. Rubin, 2002, Suitability of GPR for characterizing variably saturated sediments during transient flow, Proceedings of Ground Penetrating Radar conference (GPR 2002), Santa Barbara, CA, April 28-May 2.
  • Kowalsky, M.B., Y. Rubin and P. Dietrich, 2002, Effects of fluid distribution and subsurface structural heterogeneity on ground-penetrating radar, Extended Abstract for the 2002 International Groundwater Symposium, Berkeley, CA, March, 2002.
  • Hubbard, S., Grote, K, Kowalsky, M. and Rubin, Y., 2002, High-Resolution Estimation of Near-Surface Water Content using GPR Ground Wave Information, presented at the International Groundwater Symposium Proceedings, p. 169-172, March 25-28, Berkeley, CA.
  • Grote, K., S. Hubbard and Y. Rubin, Field-Scale Estimation of Volumetric Water Content using GPR Groundwave Techniques, Water Resour. Res., Vol. 39, No. 11, 1321, 10.1029/2003WR002045, 2003.
  • Hou, Z., S. Hubbard, and Y. Rubin, Bayesian forward and inverse modeling of vadose zone hydrogeology, submitted to Water Resour. Res., 2003.
  • Hubbard, S. Grote, K. and Rubin, Y., Estimation of near-subsurface water content using high frequency GPR ground wave data, Leading Edge of Exploration, Society of Exploration Geophysics, VJ. 21, No. 6, 2002.
  • Huisman., S. Hubbard, S., D. Redman, P. Annan, Monitoring Soil Water Content with Ground-Penetrating Radar: A Review, Vadose Zone Journal, 4(2), 2003.
  • Lunt, I., S. Hubbard, and Y. Rubin, Soil Moisture Content Estimation using Ground-Penetrating Radar Reflection Data, Submitted to Journal of Hydrology, 2003.
  • Grote, K., S. Hubbard and Y. Rubin, Field-Scale Estimation of Volumetric Water Content using GPR Groundwave Techniques, in revision for publication in Water Resources Research, original submission: May 2002.
  • Grote, K., Hubbard, S. and Rubin, 2002, Y., Monitoring spatial and temporal variations in soil water content using GPR reflection data, Leading Edge of Exploration, Society of Exploration Geophysics, v21(5), 482-485.


Progress 10/01/01 to 09/30/02

Outputs
There are currently no techniques available to yield information about water content at both the resolution and spatial coverage needed to assist in many subsurface problems, and in particular, precision vineyard management. The objectives of this research are to investigate the applicability GPR methods to provide accurate and high-resolution estimates of soil water content under real field conditions. Our research has focused on water content estimation using GPR groundwave and reflection data, and use of the estimates for predicting the soil profile evolution. This year we have completed our investigation on the accuracy and feasibility of the GPR groundwave technique, by performing detailed full field-scale studies at the Robert Mondavi Winery, near Napa, California. Detailed studies were performed to determine optimal GPR acquisition and interpretation procedures, and to assess accuracy. Investigations showed that common offset 900 MHz GPR groundwave data could yield estimates of volumetric water content with a root mean square error of 0.011 compared to gravimetric measurements. The studies showed that zone of influence of the GPR groundwave at this site was from 0-20 cm, and that soil texture greatly influenced the moisture content. The second phase of the study entailed collecting densely spaced GPR measurements using 900 and 450 MHz antennas over the entire study site. Estimates of near surface soil water content from these data revealed the effects of both seasonal precipitation and irrigation with high resolution. A persistent spatial pattern of water content was revealed by the moisture maps, which is interpreted to be influenced by soil texture. The dense GPR data were used to assess the site water content spatial correlation function and to investigate the role of farming practices, precipitation, and irrigation on water content distribution. In order to obtain information about deeper soil properties, we are also investigating the utility of GPR reflections, or events returned from soil layer interfaces, for estimating water content. To test this concept, we have developed a new study site at the Dehlinger Winery near the town of Sebastopol, California. Within this area, we are investigating the influence of natural subsurface channel feature, identified using surface 100 MHz GPR data, on water content distribution over space and time. In the last year we have collected borehole and surface GPR data grids, and have developed a preliminary site-specific petrophysical relationship. We will focus this year on quantifying the accuracy and feasibility of GPR reflection methods for deeper soil layer water content estimation. Another part of our study is to develop a method for predicting the evolution over time of soil moisture profile over the entire rooting depth, based on periodic measurements of the soil moisture conditions at the ground surface (such as from GPR data) and numerical predictions. The significance of this approach is that it will allow us to profile the soil moisture using GPR groundwave water content estimates.

Impacts
There are currently no techniques available to yield information about soil heterogeneities and water content at both the resolution and spatial coverage needed to assist in many subsurface problems, and in particular, precision vineyard management. Our research is aimed at investigating the potential of surface-based Ground Penetrating Radar (GPR) techniques for use as a water content field tool. Our research to date suggests that both GPR groundwave and reflected wave methods can yield accurate, spatially dense and non-invasive estimates of shallow water content in large-scale field applications. We have used the dense GPR estimates for investigating the impact of framing practices on the spatial variability of water content as well as the role of spatially variable soil properties on ecosystem parameter predictions, and expect that GPR-obtained information will be of use to other managed and natural ecosystem studies.

Publications

  • Grote, K., Hubbard, S. and Rubin, 2002, Y., Monitoring spatial and temporal variations in soil water content using GPR reflection data, Leading Edge of Exploration, Society of Exploration Geophysics, v21(5), 482-485.
  • Hubbard, S. Grote, K. and Rubin, Y., 2002, Estimation of near-subsurface water content using high frequency GPR ground wave data, Leading Edge of Exploration, Society of Exploration Geophysics, v21(6), 552-559.
  • Kowalsky, M.B., Rubin, Y., and Finsterele, S.A., 2002, Inversion of Hydrogeological and Time-lapsed GPR Data for Flow Parameters, AGU, 83(47), Fall Meeting Suppl., Abstract H61A-0750.
  • Zhangshuan Hou, Yoram Rubin, and Susan Hubbard, 2002, Numerical Simulation of Soil Water Content in Vadose Zone Using Constraints Provided by Geophysical Measurements, Eos. Trans. AGU, 83(47), Fall Meeting Suppl., Abstract H61A-0746.
  • Kowalsky, M.B., and Y. Rubin, 2002, Suitability of GPR for characterizing variably saturated sediments during transient flow, Proceedings of Ground Penetrating Radar conference (GPR 2002), Santa Barbara, CA, April 28-May 2.
  • Kowalsky, M.B., Y. Rubin and P. Dietrich, 2002, Effects of fluid distribution and subsurface structural heterogeneity on ground-penetrating radar, Extended Abstract for the 2002 International Groundwater Symposium, Berkeley, CA, March, 2002.
  • Hubbard, S., Grote, K, Kowalsky, M. and Rubin, Y., 2002, High-Resolution Estimation of Near-Surface Water Content using GPR Ground Wave Information, presented at the International Groundwater Symposium Proceedings, p. 169-172, March 25-28, Berkeley, CA.
  • Kowalsky, M.B., P. Dietrich, G. Teutsch, and Y. Rubin, 2001, Forward modeling of ground-penetrating radar data using digitized outcrop images and multiple scenarios of water saturation, Water Resources Research, Vol. 37, No. 6 , p. 1615.
  • Grote, K., Hubbard, S., and Rubin, Y., 2001, Soil Water Content Spatial Correlation Estimation using GPR, Eos. Trans. AGU 82(47), Fall Meet. Suppl., Abstract H31C-0247.
  • Kowalsky, M.B., and Y. Rubin, 2001, Improved characterization of the Vadose zone with time-lapsed ground-penetrating radar, EOS. Trans. AGU 82(47) Fall Meet. Suppl.
  • Grote, K., S. Hubbard and Y. Rubin, Field-Scale Estimation of Volumetric Water Content using GPR Groundwave Techniques, in revision for publication in Water Resources Research, original submission: May 2002.


Progress 10/01/00 to 09/30/01

Outputs
There are currently no techniques available to yield information about soil heterogeneities and water content at both the resolution and spatial coverage needed to assist crop development and management, and in particular, vineyard management. Our research involves investigating the potential of a geophysical technique called ground-penetrating radar (GPR) as a field-based soil moisture tool. We are performing our research at the Robert Mondavi Vineyards in Napa County, California, within a 4 acre Cabernet Sauvignon vineyard block. Our research has two main components: (1) investigating the accuracy of soil moisture content estimates obtained from GPR data, and (2) comparison of soil-based information obtained from GPR data with plant- and airborne-based measurements to assess which approach, or combination of approaches, is most useful as input to precision vineyard management decisions. Our research during this first year of the project has concentrated on optimal GPR data acquisition, processing and inversion for obtaining high-resolution information about near surface water content. We have undertaken both 'detailed studies', designed to investigate a single component of the investigation, and 'grid studies', designed to assess variations in moisture content over the entire site as a function of season. The detailed studies have included activities such as trenching, laboratory analysis of samples, and infiltration tests performed in conjunction with GPR data collection and synthetic numerical modeling. These studies have helped us to understand the near surface GPR wave propagation under a variety of hydrogeological conditions, to develop an optimal GPR data interpretation procedure, to develop site-specific petrophysical relationships, and to assess the accuracy of GPR-obtained water content estimates relative to other (conventional) measurements. Comparison of the GPR-obtained water content estimates with conventional measurements of water content at our site reveal that the GPR method provides more accurate information about moisture content than TDR techniques, and provide denser information than both TDR and gravimetric techniques in a less invasive manner. We have also collected four full 'grids' of high-resolution (900 and 450 MHz) GPR data over our study site during different times in the year. Our approach for estimating near-surface soil water content has focused so far on use of information available from ground wave GPR events. Using the approaches developed during our detailed studies, analysis of all groundwave data within the GPR grids suggest that there is a spatial persistence of the water content patterns over time, and that these patterns are influenced by soil texture. Comparisons of the water content estimates obtained from GPR with the soil texture and remote-sensing imagery suggests that the dryer areas are associated with the areas of coarser grained near-surface soils and also with the most vigorous vegetation.

Impacts
We are investigating the potential of obtaining accurate and high-resolution information about water content and soil texture using a non-invasive geophysical technique called ground penetrating radar (GPR). This information would be useful as input to precision farming practices, such as in the estimation of plantable acreage, in the design of vineyard layout and in the design of a more efficient irrigation scheme.

Publications

  • Hubbard, S., Grote, K, Kowalsky, M. and Rubin, Y., Precision Water Content Management, accepted to the International Groundwater Symposium Special Publication "Bridging the Gap between Measurements and Modeling in Heterogeneous Media", 2001.
  • Hubbard, S.S, K. Grote, M. B. Kowalski, and Y. Rubin, Investigating Temporal and Spatial Variations in Near Surface Water Content using GPR, Accepted for publication in EOS, American Geophysical Union, 2001.
  • Grote, K.G., Hubbard, S.S., and Rubin, Y., Soil Water Content Spatial Correlation using GPR, accepted for publication in EOS, Americal Geophysical Union, 2001.