claim that consequences are unforeseeABLE is bold. That would require “weather is beyond our ken, forever.”
Maniac Extreme type argument on a minor semantic point.
We can make some pretty good guesses, but right now we have no effective means to fully and accurately predict the long-term and long-distance meteorological, geological, and hydrological side effects of a project that results in a moderate-to-major change in the annual rainfall of a region. There will be consequences that we are unABLE to forsee. Some of those consequences could be large, some could be negative. Some could be both, maybe we don’t get either.
Oh, my apologies—I am happy to concede that “currently unforeseeable” is a reasonable limitation in complex systems; I hadn’t noticed that qualifier.
And, if you had asked me four years ago “Might our weather models miss some catastrophic downstream consequence, which negates the potential value of returning jungle (now pasture) back to jungle, and preventing California droughts?” I would have given it a decent chance, which would negate the more intrusive, all-or-nothing interventions.
Yet—weather modelling is improving rapidly, with neural networks. Google is able to do “now-casting”, which forecasts local weather condition at small time scales. That sort of modelling was previously out-of-bounds, because it requires much smaller & more numerous voxels and turbulence could throw everything off due to local traffic conditions or a factory being shut down for maintenance. The fact that we have now-casting, among other steady improvements, lowers my assessment of a catastrophic blunder. Especially if we roll-out in a place like California, such that we return water to its state in the 1960s, which obviously would not be catastrophically disruptive.
So, it’s true that science misses catastrophe some times, and weather is complex, while very recent improvements in modelling reduce the risk of catastrophic disruption, especially when returning water to climate-change-parched regions, recently wet.
Maniac Extreme type argument on a minor semantic point.
We can make some pretty good guesses, but right now we have no effective means to fully and accurately predict the long-term and long-distance meteorological, geological, and hydrological side effects of a project that results in a moderate-to-major change in the annual rainfall of a region. There will be consequences that we are unABLE to forsee. Some of those consequences could be large, some could be negative. Some could be both, maybe we don’t get either.
Oh, my apologies—I am happy to concede that “currently unforeseeable” is a reasonable limitation in complex systems; I hadn’t noticed that qualifier.
And, if you had asked me four years ago “Might our weather models miss some catastrophic downstream consequence, which negates the potential value of returning jungle (now pasture) back to jungle, and preventing California droughts?” I would have given it a decent chance, which would negate the more intrusive, all-or-nothing interventions.
Yet—weather modelling is improving rapidly, with neural networks. Google is able to do “now-casting”, which forecasts local weather condition at small time scales. That sort of modelling was previously out-of-bounds, because it requires much smaller & more numerous voxels and turbulence could throw everything off due to local traffic conditions or a factory being shut down for maintenance. The fact that we have now-casting, among other steady improvements, lowers my assessment of a catastrophic blunder. Especially if we roll-out in a place like California, such that we return water to its state in the 1960s, which obviously would not be catastrophically disruptive.
So, it’s true that science misses catastrophe some times, and weather is complex, while very recent improvements in modelling reduce the risk of catastrophic disruption, especially when returning water to climate-change-parched regions, recently wet.