Without access to electricity for blowers, I would be looking to the "low hanging fruit": Really dry firewood, fine split wood, crisscrossing the firewood for increased surface exposure to the draft, frequent and evenly spaced firings, monitoring stack temperature with a temperature probe, and managing natural draft for stack temperature. A natural draft arch should operate with a stack temperature in the range of 650-800 degrees. (Source: Boiling 101 from The Institure for Maple Education.)
I agree with others who suggest that you would probably see little gain in adding passive air over fire to your arch.
What you are describing is a type of non-air tight stove with passive air flow to increase efficiency and reduce pollution.
In the 1970s, air tight stoves had some popularity. We heated our house with a couple of Vermont Castings air tight stoves. You could "throttle" them back for a slow burn overnight. The down side of these air tight stoves is inefficient burn producing lots of creosote, carbon monoxide, etc. when the draft was reduced.
EPA regulations "encouraged" design and sale of stoves that are not air tight and have dampers that cannot be entirely closed and have perforated ducting over the fire that help increase more complete combustion. The design prevents complete shutting off of air without unauthorized modification of the stove.
While this technology does somewhat increase efficiency improving heat yield, the biggest "gains" are not in found in increased heat at maximum burn rates but rather more complete burning at slower burn rates.
We now heat our old farm house with a couple of these more modern stoves. Because the stoves can't be "throttled" back like the old air tight stoves, the amount of time a single firing will last is less; the design prevents a slow, smoldering fire. However, chimney maintenace is much easier, very little creosote buildup. But as far as maximum heat output at full, rip, roaring burn, not that big a difference.
Some reading on air tight vs non-air tight stoves:
http://mb-soft.com/juca/print/theory.html