Help Fill the Gaps

The map above shows a visualization of where complete eBird checklists with effort information (i.e., no casual observations) have been entered for the years of 2004-2007. There are three levels of data density: blue, red, and white. The blue contours outline the areas that contribute 42% of eBird checklists. Most of the remaining 56% of checklists are scattered throughout the red areas, but the white areas combined contribute the final 2%.

It is easy to see that the higher densities of eBird checklists closely mirror the population centers of the United States: the greater Chicago-Madison region, the San Francisco Bay area, south-coastal California, and the entire Northeast show up as blue, along with metro regions of Denver, Tucson, Albuquerque, Seattle-Tacoma, and a few other cities.

We are concerned about those white areas. These are our blind spots--areas where we simply can't tell what birds occur there since eBird data are so sparse for those states. Our map for European Starling cannot possibly be accurate, since it suggests that they are absent from numerous areas in the upper midwest, Nevada, and scattered other spots where they are as common as anywhere.

Worse still, we still have some black holes--areas where we don't even have  five complete eBird checklists. Most of northern Canada, central Mexico, Nicaragua, El Salvador, and much of South America are our biggest black holes (black holes show up as beige on eBird maps). But notice on any eBird map that there are six 100 km squares that are black holes even within the United States (in the Dakotas, Nebraska, and Montana.)

We'd love to fill in these gaps to make our eBird coverage more uniform. However, there is only one way to do that: we need eBird checklists from those regions.

How you can help

First, take a close look at the map. Do you have some old checklists lying around from a birding trip to South Dakota, west Texas, or any of these other blind spots? Take a look at the bar charts for these regions. The bar chart for the entire state of South Dakota, for example, only shows 290 species, despite the fact that around 420 species have been recorded there. Some periods, like late October, don't have sufficient data to generate the bar chart. Even if you can't contribute a full checklist with effort information, any observation from this state could add significantly to what we have in eBird. You'd be almost sure to show up on the high counts table here. If you do have data, it is never too late to enter it in eBird.

Second, do you know any birders in those states? The network of eBirders around the country is ever-growing and each of us must know someone in those white areas who birds regularly and keeps track of their sightings. If so, drop them a line and encourage them to start eBirding. If they participate in their local bird club or have a group of friends that they bird with, encourage them to pass the word.

Third, can you put us in touch with influential birders or birding groups in those states? We'd love to make more connections in these states to try to boost participation even further. If your friends include active birders in these areas who would be interested in helping us to share eBird with birders in these states, please contact us at ebird@cornell.edu.

Fourth: travel

Planning your next birding trip? Consider going somewhere unexplored!

Although southeast Arizona, south Texas, California, Florida, and Cape May are the widely touted "hotspots" of the country, there is fantastic birding to be enjoyed in the eBird blind spots as well.

Eastern Montana and North and South Dakota don't offer much in terms of topography, but these prairie states have a fantastic selection of breeding birds. Dozens of waterfowl, five species of grebes, herons, three species of rails (including Yellow), 'Western' Willets, Marbled Godwits, and Yellow-headed Blackbirds bring the prairie potholes to life. Visits in late May or September can be excellent for migrating shorebirds and concentrations of migrating warblers can be found in the towns and other isolated patches of trees. In the surrounding prairies, Upland Sandpipers, Chestnut-collared Longspurs, Sprague's Pipits, Baird's Sparrows, meadowlarks, and a variety of raptors can be found nesting. These states may be brutally cold in mid-winter, but some birders make regular pilgrimages here to enjoy the rare-but-regular Gyrfalcons, along with abundant Rough-legged Hawks, Snow Buntings, and Common Redpolls.

Nevada is an unsung hotspot for birding, especially during migration months. Although much of the central portion of the state is sagebrush desert (appropriately populated by Greater Sage-Grouse, Sage Thrashers, and Sage Sparrows), oases of trees and the occasional lakes and river are magnets for migrating birds. Eastern warblers, vireos, and flycatchers show up at favored migrant traps regularly in May and early June and again in September and October. In winter the deserts host dozens of Buteos, including Rough-legged and Ferruginous Hawks. Close to Reno, Pyramid Lake and the Lahanton Valley have well-deserved reputations as birding hotspots, and the Las Vegas area offers the birder top-notch birding at Corn Creek and Henderson, as well as other locales.

We are shocked that central Nebraska is showing up as a blind spot for us. Thousands of birders travel there every March to enjoy the hundreds of thousands of Sandhill Cranes that stage along the Platte River in central Nebraska. And it isn't just the cranes: Snow and Ross's Geese and a number of other waterfowl use the Platte River and the surrounding fields for feeding and staging on the northward migration. During spring and fall, the Rainwater Basin is renowned for its shorebird concentrations.

We could go on about west Texas, eastern Oregon, northwest Iowa, and northern Maine. None of these areas are devoid of birds and all have their own hotspots waiting to be discovered. Best of all, these are areas  where the chance for discovery is highest. Want to find a first state record? Try looking for migrant landbirds in North Dakota in fall or checking lakes and rivers in Nevada. Hoping to contribute something significant to our knowledge of birds? Any one of these areas offers many opportunities for discovery. With species like Great-tailed Grackle, Eurasian Collared-Dove, and Glossy Ibis expanding their ranges, there is always a new outpost to be discovered or a new "earliest ever" record to find.

Why it is important

At eBird we consider every single sighting to be important whether of House Sparrows in a New York backyard or the recently-described Jocotoco Antpitta in Ecuador. Every piece of information helps us to refine our understanding. But recently we have come to feel that data from these blind spots is what we need most.

We are on the cusp of being able to make some very interesting and accurate predictions about the relative abundance of birds the United States, using  field checklists contributed by eBirders. This process may shed some light on what suite of environmental factors are most important for a given species and promises to help us generate relative abundance maps for each species at any time of the year. Similar predictive maps generated from the Breeding Bird Survey and Christmas Bird Count have been only for summer and winter, respectively. We are excited about these possibilities and the best way to improve our models is to have some checklist data from the areas of low data density. We believe that birder checklist data has great value to science and conservation, and we have confidence that we'll be able to demonstrate this using eBird data. But the more data we have, the better it will be. The more you can do to supply effort information and complete checklists, the more we can do.

Below are two maps for Yellow Warblers that we have generated using these methods. The maps use actual eBird checklists to predict where Yellow Warblers will occur and when. Each surface shows the estimated relative abundance of Yellow Warbler on a specified day in 2006. The surfaces are computed by estimating abundance at 200,000 locations across the conterminous US. At each of these locations we estimate the abundance taking into account local habitat characteristics from remote sensing data summarized at 15km resolution.

The estimated abundance is standardized as the expected number of Yellow Warblers detected by a typical eBird participant who birds for 1 hour traveling a distance of 1km beginning at 6AM. This controls for variation in detection as a function of observer effort as well as variation in bird availability for detection throughout the day. Estimated abundance is then averaged within each pixel. Finally, we scale the expected counts according to the breeding season distribution, producing a annual measure of Relative Abundance that runs from a maximum of 1.0 (maximum breeding season density) to zero.

In the future, we hope to be able to offer many more of these kinds of maps, but in the meantime, we hope to collect more data from our blind spots and black holes!