CENS Environmental Solutions - Sap Flow

Sap Flow Measurements and Instruments

◊ System Description

System Description

We have had field experiece with three kinds of sap flow sensors, the heat-dissipation (or constant-heating) type and the energy-balance (or heat-balance or collar-type) probes. Below is a brief description of our installation of each type, beginning with the commercially available heat dissipation probes (Currently deployed systems - #1), then our own home-built heat dissipation probes (Currently deployed systems - #2 and Previously Deployed Systems - manzanita), and finally the commercially available heat balance probes (Previously Deployed Systems - rhododendron).

Niles Hasselquist installing probes

Currently deployed systems - #1

We are currently testing the FLgsTDP-XM1000 system from Dynamax (manual: FLGS-TDP XM1000.pdf; 275 kb).

The intent is to test their low-power (and power cycling) system as a step towards mote-based sap flow probes. Additionally, sap flow measurements in trees clustered around AMARSS #3 and #4 stations, where CO₂ flux and root growth are being monitored, would be used as a proxy for photosynthesis in on-going studies.

We are using the Dynamax system to collect 10 constantly-powered probes alongside 10 power-cycled probes in two oak trees and two Ponderosa Pine trees. Simultaneously, we are using a Dynamax voltage regulator with our own Campbell Scientific 23X datalogger to collect sap flow from 12 manzanita branches.

I would like to note that this is the easiest sap flow system to install and make work that I have ever used.

Information provided by Dynamax on their system:

This new XM1000 unit has power efficient (~90%) power regulators for setting a stable output to heat sensors, or to pulse power to sensors. The prototype unit can do both pulse and steady state sensor power at the same time for calibration purposes, and comparison between methods.

Laurel Salzman installing probes

Eric Yuen installing probes

Advanced data logger with easy programming, data storage and retrieval capabilities - The Campbell Scientific CR1000 datalogger.

2 MB data memory, for up to 100,000 data values or 200 days of hourly records.

Built-in Program computes sap flow for specific trees.

Automatic night time Zero set. Saves processing time.

Input settings for sap wood area and tree indexing by leaf or stem area.

Power down, power saving for nighttime, power reduction.

Multiplexer for expanded channels.

High-efficiency, 80%+, voltage regulators for sensor heater power. Heater voltage monitoring included.

Two AVRD voltage regulators, supply 4 different voltages,

Six open channels for weather or soil sensors. Custom programs available for engineered sensor solutions.

Eric Graham installing probes

Dynamax probes in Ponderosa Pine

The probes appear to be very similar to the kind you can construct yourself (see Home-built Sensors), except they were made from all-metal hypodermic needles (we used plastic and metal) and heavy-gage wires with industrial connectors. Better still, the probes were coated with some sort of sealant that also allowed them to be removed without un-winding the heating coils. Very nice.

One big problem at the James Reserve was intermittent power failures. Unfortunately, the sap flow system pulls power from the backup battery indefinitely, discharging the battery completely within a day or so if power has been interrupted. Once the battery is discharged below about 10 volts, most battery chargers will not resume charging even if the power is restored (as precaution against situations such as short-circuits). This leads to large data gaps if someone is not attentive. Worse, it lead to us depleting our lead-acid batteries, permamently damaging them in the process.

Our solution was to insert a relay that totally cut off power to the system if there was a blackout that lasted more than 3 seconds (using a 12 V charger, some capacitors, and a simple 12 V coil SPDT relay). Although some data would still be lost, data collection would immediately and automatically resume once the power was restored.

Manzanita with probes

All wires lead to dataloggers

Apart from data gaps due to power issues, we have now been collecting data continuously for all of Spring of 2008 and into the Summer.

Data is downloaded manually from the CR1000 that is part of the FLgsTDP-XM1000 system. Data is downloaded automatically from the CR23X, that is running in parallel, using python software running on an Epia computer that is attached to the datalogger's I/O port via a RS-232 serial cable and wirelessly to an Internet gateway.

Stunt Ranch Satellite Gateway

Currently deployed systems - #2

Home-built sap flow probes were installed at Stunt Ranch Santa Monica Mountains Reserve, CA, along with a new Atla Labs MDA 2400 sensor board (with 24 bit conversion = µVolt sensitivity) attached to a Crossbow MICA2 868, 916 MHz mote to read the thermocouple sap flow temperature differentials.

These data are being wirelessly transmitted from the mote to a Crossbow SPB400 Stargate Gateway that is attached to our current joint USC/ISI-UCLA SensorKit installation.

Accumulated data are then sent every hour via WildBlue, a satellite-based Internet service provider.

Probes were installed in a nearby oak tree and because the location is public-accessible, precautions (like making sure loops of wire were available as strain-relief in case wires were accidentally pulled on -- a good idea in any location) had to be made.

Strain-relief loops in cabling

Installing home-built sensors

Previously Deployed Systems - manzanita

With the help of three interns in the summer of 2004 (Caitlin Hamilton, Geoff Robertson, and Marina Sharifi), we built and installed sap flow probes on one large manzanita bush (Arctostaphylos glandulosa in the Ericaceae) at the James San Jacinto Mountains Reserve.

The idea was to test the viability of working with home-built sensors and to get a system ready for integration into a wireless network.

Although the sap flow system worked well, the thermocouple voltages were too small to be read by the mote-based technology that CENS was testing at that time and so even though we collected interesting data on this one individual, integration into a larger system would have to wait.

We used a Campbell Scientific CR7 datalogging system for data collection, along with an old portable computer that ran DOS. Data collection was via RS-232 and ASCII dump of data using an old terminal program by Licor (LI-1000, here: LI-1000.zip 117 kb).

Datalogging setup

GAP sensor in manzanita leaf

Sixteen gallium arsenide phosphide (GAP) photodiodes from Hamamatsu (plastic type, G1118, spec sheet here: G1115_etc.pdf 173 kb.) were installed, four in each cardinal direction, directly on the leaf surfaces. GAP sensors are sensitive to light in the range of 400-700 nm, were calibrated against a LI-COR Quantum sensor, and thus measure photosynthetic photon flux (PPF), the wavelengths of light that drive photosynthesis.

Additionally, twelve fine-gauge (0.13 mm) copper-constantan (type T) thermocouples made from wire purchased from Omega were inserted into three leaves in each of the cardinal directions, to measure leaf temperatures. Leaf temperature was used to calculate the water vapor concentration inside of the leaf for determining the driving force for transpiration, the Vapor Pressure Difference (VPD).

Previously Deployed Systems - rhododendron

We used a Campbell Scientific CR23x datalogger with collar-type sap flow probes (Dynagage) from Dynamax to measure sap flow through a spring and summer during leaf out in Rhododendron ocidentale at the James Reserve. Data were collected simultaneously with images from a mobile camera system and micrometeorological data a few meters away.

Branches were loosely tied to fence posts driven into the ground for stability (so the camera system could return to the same location and capture images of the same branches) and aluminum foil was loosely wrapped around the upper half of the sap flow collars to shield them from solar radiation. The blue ball was placed as a marker for the camera.

Rhododendron sequence