As many of you know, we have bi-directional counters that count nearly
every bee that leaves and every bee that enters a hive.  We also have
access to technology that can identify and track individual bees.  Our
Artificial Neural Networks (a form of pattern-recoginizing software) have
proven to be useful for predicting flight activity based on actual weather
files.
 
The ANNs would do even better if we had a way to dynamically measure food
resource availability.  For pollen, we use clock-driven traps - so we know
how much pollen is coming into the hive at any time of the day.  But, we
haven't figured out how to distinguish nectar gathering bees from water
carrying bees (at least not electronically).  Because we have 28 electronic
hives at four locations and intend to expand to 38-40, we are in need of an
automatic way of distinguishing between nectar and water gatherers. In
other words, we are looking for a way of modifying our bi-directional
counters to categorize bees as nectar or water carriers.
 
We have kicked around some ideas, but haven't found the magic answer.  An
alternative is to weigh the hive as an indirect measure of the food stores
and population of bees in the box.  Two years ago, we test a pressure
transducer system connected to a hydraulic base under the hives, but found
this to be too sensitive to temperature and barometric pressure changes.
We then changed to strain gauges.  The problem is sensitivity.  When we
increase the sensitivity, the hive bottoms out.  When we adjust the range
to cover the 15-30 pound range of our small test hives, we have trouble
measuring weight changes due to the coming or going of a few bees.  We have
been playing around with a two-stage system - using counter-balance weights
or springs to carrying the bulk of the hive weight and the strain gauge
suspension system to measure small changes.
 
As always, cost is a critical factor.  We can't afford to drop an expensive
electronic balance under each of 28-40 hives.
 
Any suggestions?
 
Cheers
 
Jerry