I recently acquired a bonsai and am nervous about the danger of over- or under-watering. Is it possible/feasible/worthwhile to build some sort of drip system to keep the soil at a constant humidity? Or is the moistening/drying process necessary for the health of the tree?
If you have good drainage, you don't need a drip-system. Just water it daily in growing season, and surplus water will drain from the soil. Bonsai is practiced for thousands of years, before any dripping system hit the market. If you have indoor bonsai in winter (dormancy period), you might want to keep it colder (~15 C) and water it less in this period.
Hmmm, let's imagine a technologist and an economist in conversation regarding this issue:
Technologist: We have lots of sensors available to us to provide input into a neural net which can decide whether to drip or not and how often. It's just a matter of deciding what features (inputs) to use to make that decision. We could use temperature, humidity, soil moisture level, type of plant, porosity of the potting medium, light levels, season and so on. It's all do-able these days, and we can implement it in stages, asking for recommendations from the NN, and whether the adult supervision is in agreement with the suggestion or not this then becomes a data point we can use for supervised learning. The NN can handle both the constantly moist and dry-moist-dry situations with software. Simple!
Economist: The big problem I see is the risk of things going wrong as the learning is taking place or after implementation. Bonsai plants are designed to be unique items which have longevity and high value. One bonsai can be worth thousands and we can't afford to make a mistake. Sometimes the plants are so valuable they cannot be replaced without re-starting from scratch, so insurance would be expensive. What are the chances of your sensors giving a wrong reading? For example when sensing soil moisture you are measuring electrical conductivity, and this is a function of soil type, moisture and dissolved salt levels.
Technologist: ah but you see the neural net, given enough information, can sort all of that out automatically. With smaller computers than we have today we sent people to the moon and back!
Moral: It is probably do-able, but would require a very free-draining potting soil, a special neural net for each plant and frequent supervision for the sake of peace of mind.
This can be done without any technology more sophisticated than a water source and a wick.
It is called capillary or wick irrigation and works best for plants that are not succulents or require a dry period like cyclamens.
The wick can be made of cotton or polypropylene and should be six to eight inches long. Use a knitting needle to push the wick into the root ball. Put the other end into a water container. Water to make sure that the wick is wet enough to move water up from the reservoir into the root ball.
This method ensures that the plant will never be overwatered but does need the reservoir to be topped up. When done indoors moss can get out of hand on the wick.
Products are available from many vendors. I have used this product indoors and out and it worked well.