What the average cannot see

Most days, a wadi in Oman looks like empty ground. The stillness is convincing enough to make permanence feel plausible; heat settles over gravel, the channels remain dry, and all that open space can seem to be waiting for a more useful purpose. Then rain falls somewhere in the mountains, water gathers speed, and the landscape changes in hours. The wadi has not become something else. Its other condition has simply arrived, exposing how easily we had judged its purpose by the condition we encountered most often.
I was thinking about this again in São Paulo this week while moderating a discussion on food security and resilience. Brazil is a major global agricultural powerhouse and still imports around 85% of the fertiliser it consumes. The scale is enormous, but so is the dependence on systems whose robustness is easiest to ignore while they continue to work smoothly. Our conversation kept circling back to resilience, what allows a productive system to keep moving when the conditions beneath it shift, and how abundance can depend on things beyond our immediate control.
Olivier Hamant, the French biologist, described this changing world as a passage from the era of the mean into the era of the standard deviation. His deeper argument is about robustness, about what happens when systems built for performance encounter a world that does not behave within the range they were optimised to expect. We have become extraordinarily good at planning around averages; rainfall becomes a design assumption, temperature a threshold, demand a forecast, and forecasts slowly acquire the authority of fact.
The same instinct has shaped what we remove. Inventory looks idle when shelves are full, a second supplier looks expensive when the first is reliable, spare capacity looks wasteful when demand behaves, and open land invites another use when water is absent. Wetlands, mangroves and healthy soils carry the same difficult kind of value. Their worth is easiest to overlook during ordinary years because a buffer earns its keep in the deviation; when the shipment fails, the transformer breaks or the water rises, what once looked excessive becomes the margin that absorbs the shock.
Between sessions at the Arab-Brazil economic forum, someone told me about the capybaras that wandered into a state legislature building a few weeks earlier. These heavy-barrel shaped animals walked in from the park beside it, crossed the reception area and reached the turnstiles before stopping. The final boundary between park and parliament was a machine built to count people, faithfully holding a distinction the animals themselves had never needed.
The story made me laugh, but got me thinking about my own work sitting inside this logic. I spend much of it around projects measured in decades, where forecasts become assumptions, assumptions enter models, and models eventually harden into infrastructure. The longer the horizon, the more I wonder what disappears when precision gives us confidence that the range around the forecast can be narrowed away. Ecological overshoot sharpens that unease: as natural systems are pushed beyond what they can regenerate or absorb, the margin around the forecast narrows from both directions via what we build and what the living world can carry.
The dry wadi and the flooded wadi belong to the same landscape. One is simply easier to build our expectations around. The open ground remains through another season of heat and gravel, carrying no obvious return, while somewhere beyond sight the possibility of water remains part of its design. It was never empty. We had forgotten how to value what had not yet been asked of it.