The Art and Science of Bird Surveys in the Modern Age: A new twist to distance sampling

The author collecting bird data in the field on wetland surveys using an iPad.

Since 2012, the Missouri River Bird Observatory has been working on improving bird survey methodologies and the presentation of results by developing data collection technology through ArcGIS Collector.

We’ve blogged about it before at http://mrbo.org/an-evolution-bird-survey-data-collection-using-ipads/ and presented on the topic in Sweden at the International Bird Observatory Conference http://mrbo.org/of-birds-food-and-biostatistics/ and yet, within the bird monitoring community use of modern technology to facilitate the data collection process is relatively rare.

It takes time to understand the countless, well-developed methods for conducting bird surveys and data analysis. Each approach is often tailored by region, habitat, and taxa, and information needs. Distance.org provides a simple description of Distance Sampling, one of the most commonly-used survey methodologies.

Project Distance logoFrom the Distance folks in Scotland: “Distance sampling is a widely used methodology for estimating animal density or abundance. Its name derives from the fact that the information used for inference are the recorded distances to objects of interest (usually animals) obtained by surveying lines or points. In the case of lines the perpendicular distances to detected animals are recorded, while in the case of points the radial distances from the point to detected animals are recorded. A key underlying concept is that the probability of detecting an animal decreases as its distance from the observer increases. Much of distance sampling methodology is concentrated on detection functions, which model the probability of detecting an animal, given its distance from the transect.”

Even though Distance sampling can provide incredibly robust statistics, we at MRBO have sought for and developed a way to improve data collection and provided supplemental outputs. In short, we not only keep list of birds detected and respective distances, we also map them. To be clear, traditional Distance sampling results in estimates of density (birds per/acre) and abundance. Our data collection method provides these results with the added advantage seeing exactly where the birds are seen in the landscape.

A surveyors iPad view of a line-transect (red), gridlines (blue) for further distance reference, and bird detections (yellow points).

We have been using iPads on surveys for four years to collect data. Specifically, we use ESRI‘s software and applications, Collector for ArcGIS and ArcGIS Desktop ArcMap to conduct line-transect and point-count and prepare data for Distance sampling. Our GPS-enabled iPads are prepared with aerial imagery basemaps, elements of study design (transects and survey points), as will as feature collection services (i.e. bird and habitat) that work offline in remote settings. As surveyors walk a transect line they see where they are on the aerial imagery basemap and mark all birds where they are seen or heard.

Since 2013, we have collected locations of hundreds of thousands birds using Missouri’s most imperiled habitats. Resulting density and abundance estimates as well as mapped bird locations are provided to landowners and land managers.

In addition to the statistical analysis results, mapped results can be overlaid with management and habitat information.

Screenshot of an online web app. This webapp is a simple comparison of select species distribution at St. Joe State Park.

Mapped individual bird points or heatmaps provide a compelling visualization of factors effecting shifts in occupancy from year to year. Maps can be made available via ArcGIS online in the form of webmaps or webapps. Some of our data are available in various formats at mrbo.maps.arcgis.com.

In addition to more flexible outputs, the process has considerable advantages over traditional sampling methodologies. We save an extraordinary amount of time and resources. In any given year, we cover about 950 line transects and about fifty point counts. Traditionally, this would have required the design and printing of 1,000 datasheets. Additionally, it would have entailed hours of daily data entry with countless chances of errors in transcribing the data. Data entry now takes place in seconds without human error with the click of a button on an iPad. Field work just got a lot more fun.

Distance sampling requires estimating the distance from subjects of study to the transect line or point count point location. People are notoriously bad at estimating distances precisely. With our method each bird is placed with much greater degree of precision. Survey crews can navigate and plan logistics much more efficiently without previously being oriented in an area.

Data saved on secure servers in the cloud and on multiple hard drive backups. It can seamlessly be pulled into the desktop environment and synched back to the cloud. There is tremendous flexibility in sharing data.

So, where do we go from here? Since this methodology is so new, the many outputs and further uses of the data haven’t been fully realized. One hope is that we will be able to capture distances from observer to birds rather than distance from distances of birds to transect line. This would lead to more accurate detection functions in Distance analysis. Also, when more researchers are using these methods, datasets would be easily combined. Furthermore, there are piles of data on paper, floppy disks, CD ROMs, Zip Drives, and hard drives in unusable legacy formats. This data could be digitized in varying degrees of spatial accuracy to become more useful and relevant. Perhaps, some the money saved by organizations using these methods could be spend on the task of digitizing data.

All of these ideas of where to go from here are just ideas… but isn’t how we got here to begin with?

Comments are closed.