Abstract:
This dataset comprises the 2021 Shamrock Island Nesting Survey. The colony is located in Corpus Christi Bay in southern Texas and comprises one of the largest waterbird breeding colonies in the state. The colony was surveyed with a drone, a DJI Inspire 2 paired with a zenmuse x7 camera, over the course of three days: 2021-05-25, 2021-05-27, and 2021-05-28. On 2021-05-25, the northeast portion of the colony was surveyed during two separate missions (times 1014-1152 and 1604-1801). On 2021-05-27, the southeast portion of the colony was surveyed during one mission, time 0851-1125. On 2021-05-28, the southwest, northwest, and center of the colony were surveyed during two separate missions (times 0849-1129 and 1736-2006). This dataset includes individual photos captured during the survey, orthomosaics, and total number of birds per species.
Suggested Citation:
Mirzadi, Rostam E., Kate R. Shlepr, and Dale E. Gawlik. 2021. Shamrock Island bird nesting survey using sUAS (drone) photography, 2021-05-25 to 2021-05-28. Distributed by: GRIIDC, Harte Research Institute, Texas A&M University–Corpus Christi. doi:10.7266/V4DE0SMV
Purpose:
To our knowledge, the 2021 nesting survey of the Shamrock Island waterbird colony is the first to use images taken from a drone (Small Unmanned Aircraft System- sUAS). Nest counts provided by the sUAS for most species were close to the mean for counts recorded since 2000 in the Texas Colonial Waterbird Survey database. However, the sUAS provided much higher counts than were recorded in the earlier period for the Reddish Egret, White Ibis, and Royal Tern. Our count of Royal Tern nests was about 1.5 times higher than the previous max count for the entire period 2000-2014. Counts for the other two species, one of which is of special management concern, were just below their respective max count in the earlier period, but still considerably higher than the average. None of the counts in this study were below or nearly below the min of those in the earlier period.
We believe the ability of the sUAS surveys to provide average to higher than average nest counts while causing minimal disturbance to nesting birds is strong justification for expanding the use of sUAS surveys of waterbirds on the Texas coast. If that happens, it will be useful to address two key uncertainties: one to preserve the value of the early data in the Texas Colonial Waterbird Survey and another to increase the efficiency of using sUASs for surveying large colonies. The Texas Colonial Waterbird Survey allows for nest estimates based on flight line counts and direct counts of nests, but there is little information on how nest estimates using those techniques would compare to those take from sUAS images. Thus, one key research need is to develop a correction factor for sUAS surveys that would allow for nest estimates based on that method to be comparable to others in the Texas Colonial Waterbird Survey database. This could be accomplished by conducting surveys at a subset of colony types using both methods for several years. Secondly, the sUAS surveys are probably most valuable for the largest multi-species colonies because they cause minimal disturbance while providing detailed counts. However, the manual methods currently used to count the large numbers of nesting birds could be prohibitively time consuming if sUAS methods were adopted for all large colonies. Research aimed at developing machine learning or other innovative approaches for accurately counting different species would be highly beneficial to the long-term monitoring of Texas colonial waterbirds.
Data Parameters and Units:
Data for the 2021 Shamrock Cove survey are provided in four file formats. Individual pictures captured during the survey are provided as JPEGs. All JPEGs include the date, time, and GPS coordinate associated with individual photos. Orthomosaics produced during the survey will be available through GeoTIFFS and KMZ files. GeoTIFFs are provided for those who have access to ArcMap or ArcGIS. Once GeoTIFFs are downloaded, they may be added to ArcMap or ArcGIS through the “add data” tab. The user should then right-click the multiband raster layer of each RGB composite, click Properties, click the Symbology tab, click RGB Composite, click the Stretch drop-down arrow, and select “none”. For users who do not have access to ArcMap or ArcGIS, users will be provided with KMZ files. These files may be added to Google Earth. After the user has opened Google Earth, they will click on the projects tab, click the New Project button, select Import KMZ File, select the KMZ file they want to open, and the files will automatically save to their KMZ files in Projects. All KMZ and GeoTIFF files will be set to WGS84 UTM.
Bird count data are found in Revised Shamrock Bird Count Metadata.csv and Revised Shamrock Bird Count_Tern_BLSK_Distances.csv. SP refers to the species of bird under observation; all species are listed based on their 4 letter acronym. RYTE: Royal tern; SATE: Sandwich tern; BLSK: Black skimmer. Total birds is the total number of birds, per species, that were recorded by the observer during the initial nest counts. For terns and BLSKs, this number only incudes birds which were observed sitting. All standing or flying birds were excluded from nest counts. Dis. of birds to nearest neighbor (m); sum (m) - sum of distance of birds to nearest neighbor; Mean Distance (m); Max Distance (m); # of Birds below Min.; (Total Bird count) - # of Birds Below Min.; (Total Bird count) - # of Birds Below Min.
Methods:
The sUAS, a DJI Inspire 2 paired with a Zenmuse x7 camera, was programmed to photograph the breeding colony using the third-party app DJI GS Pro. The sUAS flew at an altitude of approximately 36 m and captured images at predetermined distance intervals that allowed for 80% front and side overlap between images. The sUAS batteries were recharged from a “house” battery on the boat as they were discharged. The sUAS was launched from a boat anchored >50 m from the island and >100 m from the nearest visible nest. Because of the relatively low altitude missions were flown, the high overlap of flight transects necessary for high resolution post-processing, and strong winds, batteries were depleted quickly resulting in a shorter than expected duration of individual flights. Thus, the colony was surveyed over the course of three days: 2021-05-25, 2021-05-27, and 2021-05-28. On 2021-05-25, the northeast portion of the colony was surveyed during two separate missions (times 1014-1152 and 1604-1801). On 2021-05-27, the southeast portion of the colony was surveyed during one mission, time 0851-1125. On 2021-05-28, the southwest, northwest, and center of the colony were surveyed during two separate missions (times 0849-1129 and 1736-2006).
After finding the “Original Nesting Pair Count”, the observer went to an area of the orthomosaic image in which all birds observed were most likely sitting on nests. For RYTEs and SATEs, these birds were identified based of several criteria; 1. they were sitting, and the observer could identify that their tail was pointed up (a trait indicative of a bird sitting on a nest), 2. The observer could identify birds sitting on what appeared to be crèches (this is the area chicks move to once they have hatched until fledging), and 3. Individual birds have chicks which are visible to the observer. The observer than used the Measure feature in ArcMap to measure the distance of 10 individual birds believed to be on nests, recorded in meters, to their closest neighbor. After measuring the distance of birds to their closest neighbor, the minimum value was used as the radius in ArcMap’s Generate Near Table feature. We then applied the min nearest-neighbor distance in this feature to cull out 1 bird of a pair that were sitting close together.; “Min” is the shortest distance between neighboring birds observed from the 10 sampled birds; “mean” is the sum of the distances between neighboring birds measured for all 10 sampled birds, divided by 10; and “max” is the greatest distance between neighboring birds observed from the 10 sampled birds. For BLSKs, birds were identified based on 2 criteria; 1. a bird’s proximity to other birds (birds are evenly spaced and not very close to one another); 2. “scrapes” (areas in the sand which breeding pairs dig prior to nesting) are visible at or near the location of a sitting bird.
View Metadata