Once upon a time, I used to make my living making maps. Not making maps, really, using existing maps to collect data to build better maps. We collected data on the plant communities across national parks, to build vegetation maps that can be used to protect endangered plants, preserve critical habitat for animals, study the effects of climate change, and basically provide a baseline for any kind of ecological research that might be done in the park. I’ve worked on these projects at Denali, Lake Mead, Death Valley, and Mojave National Preserve. The process, at all of them, is pretty similar, although the temperature, scenery, and flora certainly varied. Here’s the story of how my crew mapped the plants across Death Valley National Park.
Step 1: Figure out how to hike without a trail. As you can see from the map above, Death Valley National Park covers an enormous expanse of the Mojave desert, not just the infamous valley, but also 8 mountain ranges, sand dunes, canyons, and hot springs. Only a few roads traverse the landscape, and even fewer trails. Most of the popular hiking routes wind up canyons, natural routes formed by centuries of rare, rushing water events. But, the idea of constructing a park-wide map means that we needed to explore beyond the easily accessed locations.
The first thing people learn, hiking on our crew of desert wandering botanists, is that hiking without a trail is an entirely different sport from hiking on a trail. You have to decide if you want to switch-back up the mountain slope, or attempt to go straight up the grade. On a rocky slope, you have to decide, every step, where you can place you foot and the rocks will hold up beneath you, and where your boot could slip loose. You navigate around plants, trying always to minimize your effort as you maximize your distance. With practice, after weeks in desert, all of these decisions, foot placements and routes, become routine, processed without conscious effort, so the brain is free to move on to step 2.
Step 2: Learn how to navigate across the landscape. Once you can hike without slipping or tripping, the next step is to figure out where you need to go. This is why we need maps to make maps. Death Valley covers more than 3 million acres, ranging from -282 ft in elevation in Badwater Basin to 11,000 ft at Telescope peak. Obviously, it would take way to much time and money to send botanists to every acre. Instead, we broke the park into regions, stratified those by elevation, and then generated a random sample of points to hike to and sample. Armed with a topographic map, showing the locations of the target points, a GPS containing the coordinates, and the classic compass, we had to figure out how to get where we were going.
Navigation with a GPS seems easy. It knows where you are, and where you need to be. However, the basic models that we used had no idea what the landscape looked like, GPS coordinates exist in a vacuum, not bothered with mountains or lakes that might actually be in your way. Let’s say that your GPS says that your destination is 2.2km to the NW. You can see, to the NW, a broad valley that leads to a steep ridge. Is the destination on the slope? Is it on the slope on the back side? Is it in the valley on the other, and it would be better to go around the ridge then up, over, and back down? The GPS doesn’t have these answers. You need topographic maps, a compass, and a good sense of distance, and how fast your team travels on a variety of terrains (fast in the valley around the ridge, slow climbing the steep slope) to decide how best to reach your target destination.
Step 3: Learn to identify the plants. Once the target location is reached, the data collection begins. We made lists of every plant species found in a 100² m plot, and estimated the amount of the area covered by each species in the plot. The number of species ranged from 0, in the low, low, salt and mud covered valley, to more than 30 in the diverse shrub-lands on the mid-elevation mountains slopes. In the spring, identifying plants is fun, because you can use the characteristics of the flowers and leaves to find each species in a dichotomous key. With practice, you learn the key characteristics of the common species and the patterns of relationships between related species, so that if you encounter something new, you have a head start because you can tell what family it belongs to, who it’s relatives are.
In the fall and winter, we practice forensic botany. The shrubs and trees are still easy, but the small annual plants have to be identified from the skeletons that they leave behind after they fried to a crisp in the hot summer.
Step 4: Learn to identify the plant communities. Identifying plants is great, but you can’t actually make a map with data of all of the hundreds of species we found across the park. Instead, we map plant communities, which are usually defined by the dominant species. In some ways this is simple, the way forests are usually named for the dominant tree. Some places, this is easy, most of the giant valleys are covered in a low diversity community dominated by creosote bush- a creosote community. But, what other species occur with the creosote depend on if you name it a creosote-saltbush community, or a creosote-burrobush community. Frequently, the community you are sampling fits into the pattern of 20 or 30 common community types, but occasionally, you find a rare, exciting new vegetation pattern. The data we collected is used to define the “official communities”- known as Vegetation Alliances, through statistical analysis, which figures out the significant patterns, and which species, like creosote, are the indicators of those patterns. Then, these alliances are applied to our sampling points.
Step 5: Find the patterns that control the communities across the landscape. So far, we’ve collected data at a collection of points, totaling just over 600 sampling sites. But, how do these 600 points really cover 3 million acres that we were supposed to map? At every site, after the plant data is collected, we describe the extent of the plant community on the landscape. Does the creosote cover the entire 5 miles of the valley we hiked through to get here? (Yes) Or perhaps this community of tall Antelope Bush is found only along this dry creek-bed, running in a linear flow across the slope where the occasional extra water flows. The barrel cactus are only found on the rocky slopes that are not too steep. The juniper trees only grow dense on the eastern, protected side of the mountain, and the western slope is wind-worn and sparse. These patterns, and their size, are the most important pieces of information we collected, that allow the cartographers to actually use our reports to connect communities to landscapes.
Step 6: Spend a month or entering the collected information into a giant database, checking, re-checking, writing reports, organizing photos, identifying unknown plants, trying to read your coworkers terrible handwriting or interpret stupid mistakes made by the brain after hiking for 9 hours in 110 degree heat. (Don’t worry, it’s a dry heat) I have nothing to add to this section. Blah.
Step 7: Send the results of step 6 off to the next organization in the giant NPS web. The National Parks Service Inventory and Monitoring Program is an enormous, complex program involving ecologists running statistical analysis, field crews collecting data, aerial photography, cartographers who actually draw the shapes on the maps that represent our data, a first draft map, which then gets checked by a smaller field data collect by another field crew, and finally, a published map, and data, that can be used by researchers and managers to understand and protect the plants, and the animals that depend on them, in our national parks. The whole process takes 4 or 5 years, so perhaps by 2015, I’ll be able to see our maps.
Special thanks to Scott, who hired me on to this project, Teague, Sarah, Karin, Shannon, Joslyn, Carl, Julia and Cayenne, who hiked around the desert with me, and Jeanne, who kept us on track. Can you tell I miss the desert? All photos taken by me while exploring Death Valley, with the exception of the one Lake Mead shot.