Skip to main content

Data and code from: A large-scale experiment demonstrates line marking reduces power line collision mortality for large terrestrial birds, but not bustards, in the Karoo, South Africa

Cite this dataset

Shaw, Jessica et al. (2022). Data and code from: A large-scale experiment demonstrates line marking reduces power line collision mortality for large terrestrial birds, but not bustards, in the Karoo, South Africa [Dataset]. Dryad.


Line markers are widely used to mitigate bird collisions with power lines, but few studies have robustly tested their efficacy. Power line collisions are an escalating problem for several threatened bird species endemic to southern Africa, so it is critical to know whether or not marking works to adequately manage this problem. Over 8 years, a large-scale experiment was set up on 72 of 117 km of monitored transmission power lines in the eastern Karoo, South Africa, to assess whether line markers reduce bird collision mortality, particularly for Blue Cranes Grus paradisea and Ludwig’s Bustards Neotis ludwigii. We tested the two marking devices commonly used in South Africa: bird flappers and static bird flight diverters. Using a before-after-control-impact design, we show that line marking reduced collision rates for Blue Cranes by 92% (95% CI 77-97%) and all large birds by 51% (95% CI 23-68%), but had no effect on bustards. Both marker types appeared similarly effective. Given that monitoring at this site also confirmed high levels of mortality of a range of species of conservation concern, we recommend that marking be widely installed on new power lines. However, other options need to be explored urgently to reduce collision mortality of bustards. Five bustard species were in the top ten list of most frequently found carcasses, and high collision rates of Ludwig’s Bustards (0.68 birds·km-1·year-1 uncorrected for survey biases) add to wider concerns about population level effects for this range-restricted and Endangered species.  

This dataset includes the data and R code for this journal paper.


This dataset was collected through surveys of high voltage power lines near De Aar in the Karoo, South Africa 2008-2016 to identify dead birds that had collided with the power lines. Part of the line was marked with flappers and bird flight diverters in 2011 to allow testing of whether these marking devices reduced the collision rate for the most common species.

There are 2 csv files containing data:

1. BootstrapOct2020 - this data was used to generate collision rates through bootstrapping in the pre-marking period. It is the number of collisions per span check by top 10 most commonly recovered species. The file gives line ID, powerline (DH2, HP1 or HP2), year (1 or 2), season (winter, spring, summer or autumn), survey (date), span number and then number of dead birds found of LB (Ludwig's Bustard), BC (Blue Crane), NBK (Northern Black Korhaan), WS (White Stork), Sec (Secretarybird), KK (Karoo Korhaan), PC (Pied Crow), BK (Blue Korhaan), SWG (Spur-winged Goose) and KB (Kori Bustard).

2. MarkingOct2020 - this data was used to assess marker efficacy. It is the number of collisions per section check in the reduced experiment for 4 bird groups, with the number of spans in each section check. The file gives line ID, date, season (winter, spring, summer or autumn), experiment (i.e. before or after the markers were installed), powerline (DH2, HP1 or HP2), block (2-8), span number, treatment (as either control (acontrol), flapper or bfd (bird flight diverter)), TreatSimple (as either control (acontrol) or marked), LB (number of dead Ludwig's Bustards found), BC (number of dead Blue Cranes found), BustardsOther (number of dead bustards excluding LBs found), LargeOther (number of large dead birds other than bustards or BCs found), TotalLarge (number of all dead large birds found - i.e. sum of LB, BC, BustardsOther and LargeOther) and count_of_spans (number of spans in the treatment section).

There is one R file 'CONDOR-20-044 - R code De Aar marking experiment - Oct 20' which gives the code to run all the analyses for bootstrapped collision rates and efficacy of line marking.

There is one ReadMe file containing all necessary information.

Usage notes

ReadMe file uploaded.


Eskom (South Africa)

Mohamed Bin Zayed Species Conservation Fund

University of Cape Town

Mazda Wildlife Fund

Abax Foundation

FitzPatrick Institute DST-NRF Centre of Excellence