Research Objectives

To develop methods for obtaining real-time, simultaneous field data on root/mycorrhizal dynamics in conjunction with soil respiration (CO₂), nitrogen flux, temperature and moisture. Current minirhizotron cameras are cumbersome to operate and are very time-consuming during data collection, management, and interpretation. However, they provide valuable in situ observations impossible through other means. Our goal is to provide instrumentation that will measure multiple parameters of belowground processes and transmit data via a wireless embedded network. The resulting multiple channels of information will provide a new synthesis for addressing challenges such as quantifying carbon and nutrient flux through soil over varying spatial and temporal scales.

APPROACH

Our approach is to expand upon existing minirhizotron camera and soil sensor technologies by automating and integrating data collection through wireless networks embedded in field study sites. We are proceeding in two phases: (1) installation of prototype systems in a laboratory mesocosm and in the field, composed of conventional minirhizotron tubes in conjunction with soil sensors; and (2) modification of the prototypes through conversion to automated minirhizotron cameras and addition and miniaturization of soil sensors. The first phase will allow us to "de-bug" array design with regard to field challenges (e.g., edaphic conditions, impacts from animals and weather, providing power) and experimental design (sampling scale and arrangement). The second phase will demonstrate the value of this approach to addressing ecological issues such as quantifying carbon and nutrient flux through soil over varying spatial and temporal scales.

SYSTEMS / EXPERIMENTS

A test bed minirhizotron array was developed within a soil mesocosm in a greenhouse on the UCR campus. The mesocosm was initiated with field soil containing mycorrhizal fungal spores and a seed bank of local plants. Three minirhizotron tubes, CO₂ sensors, and soil moisture psychrometers were installed to monitor belowground processes.

Deployment of a field prototype minirhizotron array began on March 19, 2004, with installation of 15 minirhizotron tubes at the San Jacinto James Reserve. Each 1 m x 10 cm tube is constructed of clear acrylic and is inscribed with a row of 1-cm2 rectangles ("frames") along the long axis. The tubes were distributed in five sets of three arrayed along an 80 m transect directly beneath the proposed prototype NIMS line. Beginning fall of 2004, we initiated development of a wireless data logging/ in situ analysis system, capable of organizing data from multiple, different sensor systems (CO₂, temperature, moisture, nitrate). In winter of 2005, we began installation of the prototype arrays into the minirhizotron tube array (Fig 25).

Figure 20 - Schematic diagram of soil sensing array and data logging to be deployed along side the minirhizotron tubes at the James Reserve

Finally, we began conceptual development of a wireless minirhizotron robotic system capable of imaging biological processes equivalent to a 400x microscope (Fig 27).

Fig 21. Functional design of automated minirhizotron cameras.

Problems Encountered:

Our first problem was that different sensors (CO₂, moisture, temperature) have different data reading systems. We needed to develop a data-logger capable of taking data from different systems, organizing it, and then transmitting wirelessly.

Second, development of the minirhizotron requires a separate, large additional funding source.

Third, obtaining the images is only the first step. A major difficulty is actually processing the large number of images obtained. For example, we already have ~100,000 images that have been analyzed by direct observation. This process is time-consuming, and subject to individual investigator biases, making comparative analyses difficult.

Finally, several of the sensors included in the array including the cameras, thermocouple psychrometers, and the CO₂ sensors require reliable power sources in order to record measurements. At present, we are able to provide power and carry data through trenched cables incorporated into the test-bed site. Ultimately, however, we will need to reduce power requirements and provide some combination of battery and solar power at the array location. The consequences of weather and hydrological conditions to survival of the sensors remains an open question which may yet present us with some challenges. Also, a major challenge will be to develop image-processing software that can automatically recognize and quantify roots and fungal structures in digital image files, filtering many gigabytes of data and saving only that which is useful. Although digital image analysis software exists, it is designed to be controlled and operated in real time by a human a laborious task that would be impossibly time-consuming given the quantity of raw image data that will be recorded. The perceptual algorithms the human brain employs to recognize biological structures of interest against a noisy background of visually similar but uninteresting patterns in the soil are not easily described or translated into computer code.

Major Accomplishments:

The greenhouse array prototype was used to initially test an array structure. The information is currently being organized into a publication. The CO₂ concentrations showed a spatial integration with the variable temperature, moisture and distribution of roots and mycorrhizae within the mesocosm. Together these gave a very complex portrait of contrasting processes. This successful testing of a prototype array of sensors in a greenhouse mesocosm has moved us to the next step of array installation and data collection at the San Jacinto Mountains James Reserve field site. Currently, the field test-bed includes five arrays of three conventional minirhizotron tubes beneath a NIMS installation. Minirhizotron images are being collected intervals of 1 to 7 days. Two sets of soil temperature, moisture, water potential, and CO₂ sensors are being installed at three depths (2, 8, and 16 cm) in close proximity to each of the five arrays. To supplement information recorded by the array sensors, we are mapping the sub-surface ecosystem using ground-penetrating radar, and we are employing soil CO₂ efflux chambers and eddy covariance measurement techniques to address the total movement of carbon into, out of, and through the soil system.

The proposal submitted to the Biocomplexity Program of the National Science Foundation titled: Automated-Minirhizotron and Arrayed Rhizosphere-Soil Sensors (A-MARSS): Designing wireless array technology to study mycorrhizal and soil ecology dynamics, by Michael Allen (PI), Edith Allen (co-PI), Michael Hamilton (co-PI), Thomas Harmon (co-PI), and James Borneman (co-PI), was funded by the NSF. This research program is interacting intimately with CENS to develop this area of study.).

Figure 28 - Data synthesis: Short-term (< 1 day) relationship between soil temperature and CO₂ concentration at three depths in mesocosm.

Figure 22 - Data synthesis: Long term (September - April) changes in mycorrhizal root tip density (tips x cm^-2) overlaid with water potential measured in April.

Figure 23 - Data synthesis: An index of mycorrhizal turnover (as change in number of root tips) as a function of average CO₂ concentration during an extended wet phase (September through December) and an extended dry phase (January through April) in the prototype mesocosm.

Future Goals and Objectives:

As image analysis challenges and solutions become defined, collaboration in development of effective algorithms and writing "expert-system" image-analysis/pattern-recognition software will proceed. As our fungal DNA sequence database expands, we will obtain the information necessary to identify those DNA oligonucleotide probes that will allow us to quickly characterize small-scale soil fungal communities, ultimately enabling us to progress with the development of DNA microarrays for "community fingerprinting." Once the AMARSS prototype has been de-bugged, we will then be prepared to address issues of scale in soil processes and will deploy refined sensor arrays in various habitats to compare ecosystem processes under differing conditions.

Figure 24 - Close-up of ectomycorrhizal root tip and hyphae recorded at the James Reserve using a conventional minirhizotron camera. Width of image is approximately three millimeters.

PEOPLE

Faculty:

Michael F. Allen
Michael P. Hamilton

Staff:

William Swenson
Petra Prouzova
Michael Taggart
Michael Wimbrow
Scott Graham

Students:

Alisha Glass
Chris Glover
Niles Hasselquist
Natasha Ly
Ayesha Sirajuddin
Rodrigo Vargas
Alysia Hamrick