How many scientists does it take to make a pizza the size of a continent?

Jeff Taylor wrote in his last post:

When I originally heard that NEON was going to be the world’s first continental-scale observatory, I was a bit confused. I had spent years working at observatories that were part of international networks dedicated to monitoring climate change, and they all worked at only local scales. How could one observatory monitor an entire continent?

Like Jeff, I've been curious about how exactly all the measurements NEON makes on small patches of earth will stretch to describe the entire U.S. I'm starting to learn a little more about this scaling business. It's absolutely crucial to NEON's mission to provide ecological data on a continental scale. Getting hundreds of types of data that'sthat are maximally useful to many science specialties and also scalable over time and space is quite challenging. Collecting data from several different chunks of space and time and integrating nicely into a single, large-scale data set requires tight collaboration between many scientists from different specialties so that their methods, execution, and analyses all produce compatible results of the highest possible scientific value. The complex challenge of scaling calls to mind the art of making a real Brooklyn-style pizza - pizza that is perfectly thin, crisp and chewy and served in enormous slices that you must fold in half to eat properly. To get that perfect crust, you need a perfect dough. The right protein content and uniform hydration makes for a ball of strong yet tender dough that stretches into a flawless, even disc when the pizza maker applies a little physics and a lot of skill to it. Missing any of those essential components, your pizza crust will be a sad, droopy mess that tears on your fingers when you peel it off the ceiling. Stretching little batches of site data to cover the country without holes calls for a keen and thorough application of statistics as well as the scientific expertise of a wide variety of NEON staff. In other words, lots of smart people have to do lots of prep work, test runs, and analyses to determine what works, what doesn't, and how NEON should operate come time when the sites are fully built and operations are in full swing. It's a bit like planning and rehearsing an extremely complicated outdoor wedding where everyone wears field gear and nobody eats cake.
For instance, while our team was deploying remote sensing equipment in airplanes over our candidate core site in Florida last year, a ground crew was slogging through sticky August heat and measuring trees along carefully planned transects in the same area. As we wrote last year, one of the goals of that pathfinder mission was to gather ground-based measurements that we could use to calibrate and check the quality of the remote sensing data gathered by our Airborne Observation Platform. Another goal was to use the data to evaluate potential scaling strategies for plant data.

In the year since the data were collected, NEON plant ecologist Courtney Meier has been working with NEON remote sensing experts on crunching the data to complete both of the described goals. He and his colleagues have conducted statistical analyses to check things like the overall quality of the data, the effect of elevation change on the ground sampling data, as well as the optimal number of sampling points and distance between them. This information helps scientists figure out how to collect data on the ground in a way that's compatible with the remote sensing data and makes the best use of time and resources. Then there's the task of formulating and evaluating methods for collecting and melding organismal-scale plant data and regional-scale plant data. At the heart of the scaling challenge is relating some of the ecologically useful plant biomass data collected on the ground - e.g., leaf area index - to a somewhat different set of measurements - e.g. canopy height and diameter - captured from several thousand feet in the air by remote sensing equipment. There aren't direct ways of making some of these ground-based measurements from the air, so the ability to relate airborne measurements to the ground-based measurements we have could extend this ground-based information to describe a much larger area than is practical for our technicians to cover in person. Like the perfect pizza dough, the ground-based data needs to take a spin in the air to really stretch out. But for the dough to be strong enough not to tear when it's stretched, it needs to be made properly with the right ingredients. In addition, the ground-based data need statistical power to hold up to stretching over a wider area, and that starts with the right data being collected the right way on the ground. The information NEON has gathered from the 2010 pathfinder mission has suggested to NEON scientists some necessary adjustments to data collection to make it all work in Florida. Similar test runs and further adjustments will be needed to cover the huge diversity of ecosystems represented by over 50 NEON sites. "We also need to interact with the science community to get feedback about what data in what format is most useful," Courtney told me.