Briefing Paper on the Need for Research into the Cumulative Impacts of Communication Towers on Migratory Birds and Other Wildlife in the United States
Division of Migratory Bird Management (DMBM), U.S. Fish & Wildlife Service – for Public Release

LAST UPDATED:  April 17, 2009        [Comm Tower Research Needs Public Briefing-2-409.doc]

ISSUE:  The number of communication towers including radio, television, cellular, microwave, emergency broadcast, national defense, and paging towers has grown exponentially in the U.S. over the past decade.  These towers present health and safety challenges for humans, but they are also a growing impact to populations of migratory birds, 4-5 million of which are conservatively estimated to die each year in tower and guy-wire collisions (Manville 2005, 2009).  Virtually unknown, however, are the potential effects of non-ionizing, non-thermal tower radiation on avifauna, including at extremely low radiation levels, far below maximum safe  exposure levels previously determined for humans.  

This briefing paper addresses the need to cumulatively assess the impacts of communication towers on migratory birds both from collisions and radiation, especially neotropical migratory songbirds that are most impacted (Shire et al. 2000).  The paper discusses some suggested research protocols needed to conduct a nationwide cumulative impacts analysis that would assess effects of tower collisions and radiation on avifauna and on other wildlife pollinators including bats and bees.  

BACKGROUND  
 
Light Attraction to Birds in Inclement Weather   
Beginning with the earliest reported bird-tower kill in the U.S. (in September 1948 at a 137-m [450-ft] radio tower in Baltimore, MD [Aronoff 1949]), the nighttime attraction of lighting during inclement weather has proved to be a key liability for birds.  However, much of the past research focused on carcass collections that were not necessarily correlated to nighttime lighting or to weather events.  For example, the first long-term study of the impact of a television tower on birds began in 1955 by the Tall Timbers Research Station in FL.  After the first 25 years of the study, 42,384 birds representing 189 species were tallied (Crawford and Engstrom 2001).  Kemper (1996) reported collecting more than 12,000 birds killed in inclement weather on one night at a television tower in Eau Clair, WI.  Manville (2005, 2007) provided additional details of documented bird-tower collision studies in the U.S., especially in regard to lighting and weather events.  

Recently, Gehring et al. (2006, 2009) reported where red, steady-burning lights were extinguished allowing only flashing or strobe lights to persist on towers, the lighting change-out resulted in up to a 71% reduction in avian collision mortality at towers in MI.  In a short-term study, Evans et al. (2007) looked at lighting attraction at ground level in complete cloud cover, but found that neither red, steady-burning nor red flashing lights induced bird aggregation.  They hypothesized that the disorientation to red light only occurs if birds are actively using magnetoreception and the red light creates an imbalance in the magnetoreception mechanism.  Additional studies are underway to better understand the mechanisms of lighting attraction.  

    Published research protocols developed to count and estimate bird-tower kills have been developed (e.g., Avery et al. 1978, Manville 2002, Derby et al. 2002, and Gehring et al. 2009) and will be briefly reviewed below for use in future cumulative effects assessments for both collision and radiation studies.    

Potential Radiation Impacts to Birds  
In 2002, T. Litovitz (Catholic University, pers. comm.; DiCarlo et al. 2002) raised troubling concerns about the impacts of low-level, non-thermal radiation from the standard 915 MHz cell phone frequency on domestic chicken embryos under laboratory conditions.  Litovitz noted deformities, including some deaths of the embryos subjected to hypoxic conditions under extremely low radiation doses .  

    Preliminary research on wild birds at cellular phone tower sites in Valladolid, Spain, showed strong negative correlations between levels of tower-emitted microwave radiation and bird breeding, nesting, and roosting in the vicinity of the electromagnetic fields (Balmori 2003).  Birds had historically been documented to roost and nest in these areas.  House Sparrows, White Storks, Rock Doves, Magpies, Collared Doves, and other species exhibited nest and site abandonment, plumage deterioration, locomotion problems, and even death among some birds found close to cellular phone antennas.  Balmori did not observe these symptoms prior to construction of the cell phone towers.  Balmori (2004, 2005) noted that the White Stork appeared most heavily impacted by the tower radiation during the 2002-2004 nesting season in Spain.  Manville (2005) reported Balmori’s (2003) preliminary results, and raised concerns of similar events in the U.S.  

    Everaert and Bauwens (2007) found strong negative correlations between the amount of radiation presence (both in the 900 and 1800 MHz frequency bands) and the presence of male House Sparrows.  In areas with high electric field strength values, fewer House Sparrow males were observed.  Everaert and Bauwens’ preliminary conclusion, long-term exposure to higher radiation levels was affecting bird abundance or bird behavior in this species.  Balmori and Hallberg (2007) reported similar declines in House Sparrows directly correlated with levels of electromagnetic radiation in Valladolid, Spain.

    Of concern to DMBM are the potential impacts of radiation on bird populations.  Beason and Semm (2002) tested neural responses of Zebra Finches to 900 MHz radiation under laboratory conditions and showed that 76% of the neurons responded by 3.5-times more firings.  No studies have yet been conducted in the U.S. on radiation impacts to wild bird populations.  Magnetite, a mineral highly sensitive to electromagnetic frequencies (EMFs), has been discovered in human, bird, and fish brains.  It has been suggested that radio frequency radiation (RF) may be acting as an attractant to birds since their eye, beak and brain tissues are loaded with magnetite, a mineral highly sensitive to magnetic fields that birds use for navigation (Ritz et al. 2004, R. Beason cited in Levitt and Morrow 2007).  Communication tower radiation in the U.S. may already be impacting breeding and migrating populations of birds, bees, and other wildlife, based on research conducted in Europe.  It is therefore important to gain a far better understanding of the suspected impacts of radiation on birds and other wildlife, particularly if those suspected impacts are having effects on species at the population level.                

Potential Radiation Effects on Other Pollinators  
Radiation has also been implicated in effects on domestic honeybees, pollinators whose numbers have recently been declining due to “colony collapse disorder” (CCD) by 60% at U.S. West Coast apiaries and 70% along the East Coast (Cane and Tepedino 2001).  CCD is being documented in Greece, Italy, Germany, Portugal, Spain, and Switzerland.  One theory regarding bee declines proposes that radiation from mobile phone antennas is interfering with bee navigational systems.  Studies performed in Europe have documented navigational disorientation, lower honey production, and decreased bee survivorship (Harst et al. 2006, Kimmel et al. 2006, Bowling 2007).  This research needs further replication and scientific review, including in North America.  Because pollinators, including birds, bees, and bats, play a fundamental role in food security (33% of our fruits and vegetables would not exist without pollinators visiting flowers [Kevan and Phillips 2001]), as pollinator numbers decline, the price of groceries goes up.

Harst et al. (2006) performed a pilot study on honeybees testing the effects of non-thermal, high frequency electromagnetic radiation on beehive weight and flight return behavior.  They found that of 28 unexposed bees released 800 m (2,616 ft) from each of 2 hives, 16 and 17 bees returned in 28 and 32 minutes, respectively, to hives.  At the 1900 MHz continuously-exposed hives, 6 bees returned to 1 hive in 38 minutes while no bees returned to the other hive.  In exposed hives, bees constructed 21% fewer cells in the hive frames after 9 days than those unexposed.  Harst et al. selected honeybees for study since they are good bio-indicators of environmental health and possibly of “electrosmog.” Because of some concerns raised regarding the methods used to conduct the Harst et al.(2006) study, specifically the placement of the antenna where bees could contact it (i.e., potentially a bias), the experimental methods need to be redesigned and the studies retested to better elucidate and fine tune the impacts of radiation.  The results, while preliminary however, are troubling.    Kimmel et al. (2006) performed field experiments on honeybees under conditions nearly identical to the Harst et al. (2006) protocol except that bees were stunned with CO2 and released simultaneously 500 m (1,635 ft) from the hives.  However, in one of their experimental groups, they shielded the radiation source and antenna in a reed and clay box to address potential biases raised in the Harst et al. study.  Sixteen total hives were tested, 8 of which were irradiated.  After 45 minutes when the observations were terminated, 39.7% of the non-irradiated bees had returned to their hives while only 7.3% of the irradiated bees had.  

    
RESEARCH DISCUSSION  
If communication tower collisions are killing 4-5 million or more birds per year in the U.S. due to collisions, what impact – if any – might radiation have on avifauna?  Bees?  Other wildlife?  We simply do not know.  In 2000, the Communication Tower Working Group (chaired by DMBM/Manville) developed a nationwide tower research protocol that would assess cumulative impacts from tower collisions nationwide, suggesting the use of some 250 towers of different height, lighting, and support categories.  The preliminary cost estimate for a 3-year study was $15 million.  No funding was ever acquired and the collision study has not yet been conducted.  

    The proposed 2000 study was to focus on the collision impacts of communication towers to birds during spring and fall migrations, but the same types of mortality monitoring could be conducted during the late spring/summer breeding seasons, looking particularly for evidence of injury and death to breeding birds in close proximity to communication towers.  Radiation levels would need to be measured at the tower sites and nests adjacent to the towers during nesting activity, and bird behavior would also need to be monitored throughout the breeding season.  Laboratory necropsies would need to be performed on birds and other wildlife suspected of impacts from radiation to better understand what caused their deaths and to verify that they did not die from blunt force trauma from tower or wire collisions.  Pre-construction studies should be performed to assess habitat use by breeding and resident avifauna.  Post-construction studies should assess site abandonment, development of deformities, injuries, and deaths.  A careful review of the protocols developed by Balmori (2004, 2005), Balmori and Hallberg (2007), Everaert and Bauwens (2007), and others is critical because similar studies should be performed in the U.S.    
                                                                  

METHODS FOR ASSESSING AVIAN COLLISION MORTALITY

    Methods for Assessing Tall Tower Mortality    
Bird strike mortality studies at “tall”  communication towers conducted previous to research performed by Avery et al. (1978) indicated that most dead birds were found within 60 m (197 ft) of the central communication tower structure.  Avery et al. assessed songbird mortality at a 369-m (1,210-ft) Omega Loran U.S. Coast Guard tower in ND.  Based on daily monitoring during 3 fall and 2 spring migration seasons, 63% of the birds they found dead or injured at this tower were within 92 m (300 ft) of the tower.  Avery et al. placed tagged bird carcasses (e.g., House Sparrows and European Starlings) in catchment nets and on non-netted habitats (e.g., gravel pads, roads, and marshy plots) to assess persistence and scavenging/predation loss.  They completely examined the inner 46-m (150-ft) radius of the tower (concentric circle designated “A”) for bird carcasses, including both the areas covered with catchment nets and the non-netted areas.  Placing tagged carcasses in random search plots, which are then found or not found and/or removed or not removed, helps determine biases (Erickson et al. 1999).  However, there are inherent problems associated with using tagged bird carcasses, including the attraction of predators, cost, availability, and adequate sample size (D. Strickland, WEST Inc., pers. comm.).  

    In addition to the total area assessed during this study (168 ha [415 ac]), for the remainder of the search area, Avery et al.(1978) divided the habitat into concentric circles of radii 92 m (designated “B”; 303 ft) , 183 m (C; 600 ft), and 731 m (D; 2,398 ft), respectively.  Two compass lines (north-south and east-west) divided B, C, and D into 12 substrata beyond the inner core.  In each of the substratum, 2 net catchment sampling plots, 12.4 m (41 ft) on a side, were randomly selected.  Nylon netting suspended on steel frames 1.5 m (5 ft) high, with the net’s center anchored to the ground, was utilized.  See Manville (2002) beyond for additional net details.         

    Sampling nets were demonstrated by Avery et al.(1978) to be highly effective in preventing losses to scavengers and predators; none of 33 of the test birds placed in nets during the Avery et al. study were taken during the first night, but 12 of 69 test birds placed on non-netted gravel sampling plots were taken during the same period.  During the Avery et al. study, dead bird searches were made daily at dawn during the peak of songbird migration.  In a study at a Tallahassee, FL, television tower – where sampling nets were not used – scavenging was considerably higher; only 10 of 157 birds were left undisturbed after one night (i.e., 93.6% scavenging; Crawford 1971).  

    Homan et al. (2001) placed carcasses of House Sparrows in dense vegetation, comparing searcher efficiencies of humans and canines.  The dogs received no special training in carcass searching.  Thirty-six trials were conducted in 5 x 40–m (16 x 131-ft) study plots.  Humans found 45% of the carcasses while dogs found 92%.  The ratio of recovered to missed carcasses was approximately 12:1 for dogs and 1:1 for humans, making dogs much more efficient in finding carcasses.  Searcher efficiencies were not improved but remained similar when testing residual cover (April searches) versus new growth cover (August searches).  Because the protocol in the Homan et al. study improved quantitative and qualitative assessments, it provides considerable promise for the research initiatives being proposed in this briefing paper.      
      
    Arnett (2006) further tested the dog-search protocols of Homan et al. (2001) and others, assessing the abilities of dog-handler teams to recover dead bats at 2 commercial wind turbine facilities.  Dogs found 71% of the bats placed during searcher-efficiency trials at Mountaineer, WV, and 81% of those at Meyersdale, PA, while human searchers found only 42% and 14% of the carcasses, respectively.  Both dogs and humans found a high proportion of the trial bats within 10 m (33 ft) of the turbine tower, usually in open ground (88% and 75%, respectively).  During a 6-day fatality search trial at 5 Mountaineer turbines, dog-handler teams found 45 carcasses while human searchers during the same period found only 19 (42%).  As vegetation height and density increased, humans found fewer carcasses while dog-handler team searcher efficiencies remained high.  Arnett’s (2006) study further reinforces the hypothesis that use of dogs greatly improves efficiencies in finding dead bats very similar to what Homan et al. (2001) found for locating passerines.  Dog use should be given serious consideration in conducting bird and bat mortality studies at telecommunications towers.

    From 2003 through 2005, Gehring et al. (2006, 2009) studied 24 tall communication towers in MI.  They used flagged, straight-line transects, each technician walking at a rate of 45-60 m (147-196 ft) per minute and searching for carcasses within 5 m (16 ft) on either side of each transect, as suggested by Erickson et al. (2003).  The transects covered a circular area under each tower with a radius equal to 90% the height of the tower.  The straight line transects were much easier to navigate than were circular transects (J. Gehring, Michigan Natural Features Inventory, pers. comm.).  Due to dense vegetation, observer fatigue, human error, scavenging by predators, and crippling loss of birds and bats that may have escaped the detection area, Gehring et al. tested each technician’s observer detection rate and rate of carcass removal.  Ten bird carcasses of predominately Brown-headed Cowbirds, with painted plumage to simulate fall song bird migration plumage, were placed once each field season within each study plot to assess observer efficiencies.  Likewise, 10-15 predominately Brown-headed Cowbirds were placed by each technician at the edge of designated tower search area to monitor the daily removal of carcasses by scavengers.  These carcasses were not painted to avoid placing any foreign scent on them.  No catchment nets were used in this study.            

Methods for Assessing Short Tower Mortality   
Manville (2002) developed a protocol for the U.S. Forest Service (USFS) to study the effects of cellular telecommunications towers on birds and bats, recommending use of elevated catchment nets for a Coconino, Kaibab, and Prescott National Forest study in AZ.  Modifying the Avery et al. (1978) search protocol, Manville suggested use of 1.9-cm (0.75-in) mesh knitted polyethylene nets, 15 x 15 m (50 x 50 ft) in size, suspended 1.5 m (5 ft) above ground, with 8 gauge monofilament nylon line attached around the periphery of the entire net, supported with 2-m-long (6.5-ft) steel angle posts driven into the ground and spaced every 2-3 m (7-10 ft) apart.  He recommended pulling the center of each net close to the ground, securing with monofilament to a cinder block, thus creating a downslope gradient from the edge of the net to its center so a carcass landing in the net would tend not to be blown from the netting edge to the ground by a strong wind.  He did not recommend using a wooden lip on the net’s edges as Avery et al. (1978) had suggested.  Materials for each net were estimated to cost $320 (Avery and Beason 2000).  

Manville (2002) postulated that use of elevated catchment nets would make finding dead birds killed by tower strikes more reliable, especially under variable habitat conditions (e.g., unsuitable substrate for searching, tall grass, shrubs, roots, boulders, or trees).  Manville recommended breaking down the tower’s circumference into 3, 120o arcs, then breaking the study plot into 2 concentric circles.  The radius of the first circle from the tower’s center was 30 m (100 ft) and nets were to be randomly deployed to cover 24% of the total area of that concentric circle, 1 net randomly placed in each 120o arc.  For the second concentric circle (30-60 m in radius from the center [100-197 ft]), nets were placed randomly in 8% of the total area, 1 net randomly placed in each of the 3 arcs.

Manville (2002) did not recommend using tagged bird carcasses in the AZ study because he believed that double sampling would address sampling efficiency biases.  Double sampling involves (1) net sampling, allowing for an estimate of the number of carcasses that fall beneath each tower and are relatively unbiased for searcher efficiency and carcass removal, and (2) ground sampling where biases are inherent.  For short towers, he recommended the entire area the radius of the tower height be completely searched (including under the nets) at dawn each day during the migration season and once weekly during the breeding season.  Net sampling allows for adjustment of the ground sampling estimates that would correct for carcass removal and searcher efficiency bias based on the relative difference of the number of carcasses found using the 2 sampling methods at each communication tower studied.  

Manville (2002) indicated that the probability of catching a bird in a net would change with increased distance from the tower (i.e., birds my fly or be carried by the wind for a distance before dying).  He suggested that if there is a bias because birds tend to die greater than 30 m (100 ft) from a short tower, probabilities can be determined by searching strip transects that radiate from a tower.  He recommended using a transect 1.5- 2 times the height of the tower, 15 m (50 ft) wide, placed on a randomly selected compass line.  Carcass searches within the transect should help to estimate the area that should be sampled by nets, develop a correction factor outside the radius of the area sampled by the nets, and improve the correction factor for ground surveys conducted exclusive of the net surveys.  Manville suggested this transect survey be conducted at least once per week, preferably in the early morning hours, during both migration and breeding seasons.  With the recent use of trained dogs to detect and locate dead and injured birds and bats, where dogs have been shown to be at least 50% more effective in finding carcasses, dog use should be considered a viable monitoring alternative (E. Arnett, Bat Conservation International, pers. comm., Homan et al. 2001, Arnett 2006).              

Derby et al. (2002) modified the Manville (2002) protocol to conduct the cellular telecommunications tower study in AZ for the USFS.  There, 6 of the 7 cell towers were surrounded by 3-m (10 ft) walls, 29 m (95-ft) long on each side. The walled square was divided into 4 equal blocks, and within 1 of these blocks a 12 x 12-m (40 x 40-ft) nylon mesh net was randomly placed based on net specifications recommended by Manville (2002) but placed > 3 m (10 ft) above the ground to allow company personnel to perform maintenance on the sites.  Outside the walled compounds, Derby et al. used 4, 6 x 6-m (20 x 20-ft) nets, 3 of the nets randomly set outside the wall to a distance of 30.5 m  (100 ft) from the tower, and the 4th net randomly placed in the band from 31 to 61 m  (100-200 ft) from the tower.  Inside the walled compound the entire area was searched by walking transects 6 m (20 ft) apart (3 m [10 ft] search width).  The surveys were performed at dawn 4 times per week during peak songbird migration.

Derby et al. (2002) also recommended using straight line transects, 4 oriented perpendicular to the walls, and 4 diagonal from the corners of the wall – representing the “spokes of a wheel.”  Each transect was 61 m (200 ft) long, and 6-m (20 ft) wide.  Because the Derby et al. protocol also used double sampling, no tagged carcasses were used in their study.  

Both Manville (2002) and Derby et al. (2002) recommended daily searches of all electrical wiring to assess for electrocution and wire collision mortality.        

Homan et al. (2001) used Labrador retrievers and a Chesapeake Bay retriever to search 6 plots, 5 x 40 m (16 x 131 ft) in size, delineated by flagging, to detect 8 thawed House Sparrow carcasses randomly thrown in each of the plots from 1 m (3 ft) outside the plot, allowing the human or human-dog team to search each plot for 10 minutes.  Dogs were kept on 5-m (16-ft) leashes during searches.  Humans were active searchers when using the dogs.  Searches were not conducted during steady rain or when winds were > 32 km/hr (20 mph).  The technique with leashed dogs could easily be used to survey both tall and short tower plots, based on the protocols previously recommended.  With the dogs confined to leashes, additional training would be unnecessary.

Arnett (2006) used 2 trained chocolate Labrador retrievers to locate test bat carcasses of different species and in different stages of decomposition at commercial wind turbine facilities on the Appalachian Mountain front in PA and WV.  His dogs were trained in basic obedience, “quartering” (i.e., systematically searching back and forth in a 10-m-wide [33 ft] transect), and blind retrieval handling skills.  The dogs were trained with dead bats 7 days prior to field trials.  When a dog found a test bat, the dog was rewarded with a food treat if it performed the task of finding the bat, sitting or stopping movement when given a whistle command to do so, and leaving the carcass undisturbed.  Arnett walked the transect lines at a rate similar to that of humans (i.e., approximately 13-25 m/min [43-82 ft/min]) while the dogs were allowed to quarter the entire width of the transect (5 m [16 ft] on either side of the center line).  While this technique was tested on bats, it also shows great promise for use on birds.  Dogs would require additional training, but unlike the Homan et al. (2001) technique, they would not need to be leashed.  The Arnette technique also shows great promise for use at both tall and short communication towers to locate dead birds and bats.   


METHODS FOR ASSESSING RADIATION IMPACTS TO BIRDS

Methods for Assessing Radiation Impacts at Tall Towers
At present, radiation studies at tall towers in Europe have not yet been conducted since the impacts to birds and other wildlife have been documented at short, cellular communication towers.  The methods suggested below for short tower radiation studies should also be applicable to future tall tower radiation studies.

Methods for Assessing Radiation Impacts at Short Towers
Balmori (2005) selected 60 nests of White Storks in Valladolid, Spain, to monitor breeding success, visiting each nest from May to June 2003, taking care to select nests with similar characteristics located on rooftops.  Tree nests were not studied.  Nests were selected based on very high (N=30) or very low (N=30) exposure levels of electromagnetic radiation, depending on the distances nests were located from the cell towers.  Thirty nests were within 200 m (656 ft) of the towers, while the remaining 30 were located > 300 m (981 ft) beyond any tower.  Chick productivity was closely observed.  Electric field intensities (radiofrequencies and microwave radiation) were measured using a unidirectional antenna and portable broadband electric field meter set at 10% sensitivity.  Between February 2003 and June 2004, 25 visits were made to nests located within 100 m (327 ft) of 1 or several cell phone towers to observe bird behavior.  The visits were made during all phases of breeding, from nest construction until Stork fledging.  RFs and EMFs were also measured at all nest sites using a unidirectional antenna and field meter.

Balmori and Hallberg (2007) studied the urban decline of House Sparrows in Valladolid, Spain, since this species is in significant decline in the United Kingdom and western Europe, and because it usually lives in urban environments, where electromagnetic contamination is higher.  They felt it would be a good biological indicator for detecting the effects of radiation.  Forty visits, approximately 1 per month were made between October 2002 and May 2006, and were performed at each of 30 point transect locations (i.e., point counts, the protocol recommended by Bibby et al. 2000) between 7 a.m. and 10:00 a.m. by the same ornithologist following the same protocol.  At each transect site, all sparrows heard and seen were counted, without differentiating birds by sex and age, and radio frequencies and levels of microwave radiation were recorded using a unidirectional antenna and a portable broadband electric field meter set at 10% sensitivity.  Bird densities from each point were calculated based on the number of sparrows per hectare.

Everaert and Bauwens (2007) counted male House Sparrows during the breeding season at 150 point locations (Bibby et al. 2000) in 6 residential districts in Belgium, each point location situated at variable distances (mean= 352 m [1,151 ft]; range= 91- 903 m [298- 2,953 ft]) from nearby cell phone antenna towers.  Point counts were conducted for 5 minutes, all male House Sparrows heard singing or visible within 30 m (98 ft) were counted, counts occurred between 7 a.m. and 11:00 a.m. when males were most active, and counts were conducted only during favorable weather conditions.  Electric field strengths at 900 MHz and 1800 MHz were measured for 2 minutes at each frequency using a portable calibrated high-frequency spectrum analyzer with a calibrated EMC directional antenna.  To measure maximum radiation values, the EMC antenna was rotated in all directions.


METHODS FOR ASSESSING RADIATION IMPACTS TO BEES

Methods for Assessing Radiation Impacts to Bees
Harst et al. (2006) exposed 4 beehives to 1900 MHz radiation from an antenna placed at the bottom of each hive immediately under the honeycombs, while they left 4 hives unexposed.  Each of the 8 colonies contained approximately 8,000 bees.  They were set up in a row, with a block of 4 hives equipped with DECT (Digital European Cordless Telecommunications) stations on the bottom of each hive.  Metal lattices were installed between the exposed hives to avoid possible effects to the non-exposed control group.  The average transmitting power per station was 10 mW, with peak power at 250 mW.  The sending signal was frequency modulated and pulsed with a pulsing frequency of 100 Hz.  A transparent 10 cm (4 in) plastic tube with a diameter of 4 cm (1.6 in) was mounted at the entrance of each hive to collect single bees and watch them return later to the hives. Twenty-five bees from each hive were randomly selected, stunned in a cooling box, marked with a marker dot on the thorax, and released 800 m (2,616 ft) away from the hives.  All marked bees were released simultaneously and were timed from the moment of their release.  Return times were noted as the bees each entered the plastic tubes, with the observation lasting 45 minutes.  Any bees returning after 45 minutes were disregarded.  Bees were able to touch the radiation sending antenna within the hive.  Some have asserted that the antenna placement may have resulted in a behavioral bias in regard to bee response, raising a legitimate concern about the methods used to test bee response to radiation in this experiment.

Harst et al. (2006) also studied the effects of radiation on bee building behavior using the protocol discussed above.  They photographically documented change in honeycomb area, and measured development of honeycomb weight for each hive.  Sixteen colonies were selected for this experiment, 8 of which were irradiated, all aligned in a row.  At the beginning of the experiment, the empty honeycomb frames were weighed, the hives were filled with bees (400 g [14 ounces]), and provided 250 ml (0.26 quart) food.  Bees were fed 2 more times during the 9-day experiment.  The honeycombs were photographed each day.  The placement of the sending antenna, as previously suggested, may have altered bee behavior and hive productivity.

Kimmel et al. (2006) tested 16 bee colonies, 8 of which were irradiated.  The experiment was nearly identical to that utilized by Harst et al. (2006) except that the sending antenna in 1 experimental group was shielded in a reed and clay box to address concerns about behavioral biases raised in the Harst et al. study.  Bees were paralyzed using CO2 instead of cold and were simultaneously released 500 m (1,635 ft) from the hives instead of 800 m (2,616 ft).   


RESEARCH RECOMMENDATIONS FOR ASSESSING AVIAN COLLISION IMPACTS

Tall Tower Collision Research Recommendations    
    We recommend using either the Avery et al. (1978) or the Gehring et al. (2006, 2009) protocol for tall tower collision studies, depending on the feasibility and availability of catchment nets and dead bird carcasses.  Avery et al. provided the opportunity to use catchment nets, testing searcher efficiency and carcass removal by placing test carcasses on site (in nets and on the ground).  The protocol presumes that the majority of carcasses will be found within a certain distance of the tower’s base.  The protocol has particular utility for studying very tall towers, especially where terrain around the structures is highly variable and difficult to traverse.  It can be used as a standing protocol, or modified as a hybrid based on combining other techniques suggested within this paper such as the use of dogs (Homan et al. 2001, Arnett 2006).  Dogs have tremendous promise for both tall and short tower studies.  If trained hunting dogs are used, then the Arnett (2006) protocol is an excellent tool since the dogs can be used off-leash.  However, if untrained hunting dogs are available, then the Homan et al. (2001) protocol using leashed dogs is an excellent option.     

    Gehring et al. (2006, 2009) also successfully assessed mortality at tall towers, but catchment nets were not deployed in this study.  Due in part to timing, budget constraints, and number of towers studied, this protocol has significant utility where many towers need to be studied.  It could also be modified by using trained dogs or incorporating catchment nets.  

    The statistical designs for both short and tall tower studies – both for assessing collisions and radiation impacts, should be worked out with qualified biometricians.  Both the USFWS and the USGS/Biological Resources Discipline (BRD) have well qualified statistical expertise.  They should be consulted early in the development of a proposed study.

In both short and tall tower studies, data collection must include all of the following:  time of day each tower is examined, time spent searching each site, time since the last search, and weather conditions, particularly inclement weather.  Weather data should include the previous night’s temperature, wind, cloud cover (clear if < 10% cover, partly cloudy 10-90% cover, or overcast > 90% cover), barometric pressure, rainfall, fog, obscuration, and other relevant weather conditions (Derby et al. 2002).  

When bird and bat carcasses, and injured vertebrates are found, regardless of the sampling method, data must include tower identification number, name of species (if known), date of collection, closest transect, distance from the tower, azimuth to the tower, exact mapped location (GPS coordinates are very helpful), estimated number of days since death/injury, body condition, probable cause of death, and evidence of scavenging.  The carcass is to be collected, numbered, and saved to be used in other investigations (Gehring et al. 2009) for which a Federal and possibly state salvage permit will be required (Manville 2002).  

Short Tower Collision Research Recommendations
Depending on the availability and utility of catchment nets and the layout of the tower site, we recommend using either the Manville (2002), the Derby et al. (2002), Homan et al. (2001), or the Arnett (2006) protocols – the latter 2 with greatly improved searcher efficiency, or a hybrid of these methodologies.  Manville (2002) suggested using elevated catchment nets, but due to double sampling, he did not recommend using tagged bird carcasses.  He also recommended using random transects to adjust for biases.  

    Derby et al. (2002) modified the Manville (2002) protocol, specifically in regard to challenges created by the tower study site in AZ.  A randomly-placed catchment net was used within the walled enclosure of each of the sites, and the entire area within the walled compound (ground and net) was searched.  Four randomly placed catchment nets were also utilized beyond the walls. Due to double sampling, no tagged bird carcasses were utilized.  The protocol could be used as a free-standing technique but should be searched daily during the entire peak of bird migration.

    
RESEARCH RECOMMENDATIONS FOR ASSESSING RADIATION IMPACTS TO BIRDS

Tall Tower Radiation Research Recommendations
For both short and tall tower studies, any nests close to a tower should be noted, with its GPS coordinates recorded.  Breeding, nest success, and survivorship should be monitored, where possible.  How birds use their habitats for breeding and residence should be noted, including any issues of site abandonment, egg and clutch failure, development of deformities, injuries, and deaths.

For both short and tall tower studies, where birds appear to be injured or killed by radiation, proximity of the bird/carcass to known nest or roost sites and towers should be noted.  Radiation levels at the tower, carcass site, and the nest site should be recorded.  Any abnormal behaviors should also be described.  Laboratory necropsies should be performed on birds and other wildlife suspected of impacts from radiation to better understand what caused their deaths and to verify that they did not die from blunt force trauma due to collisions.  Tower and ambient radiation should be measured using equipment and techniques suggested by Harst et al. (2006) and Kimmel et al. (2006), or variations of equipment and methods available in the U.S.  See the methods section of this paper for specifics.

Where carcass counts need to be assessed at specific tall towers, we suggest using the tall tower collision mortality protocols, discussed above in the methods section of this paper.

Short Tower Radiation Research Recommendations
    Depending on the avian species being studied, we recommend using the Balmori (2005) protocol for assessing potential impacts to colonial nesting species such as herons and egrets.  Where passerines are to be studied, we suggest the use of the Everaert and Bauwens (2007) and Balmori and Hallberg (2007) protocols for assessing potential impacts.  Refer to the methods section above for specific details.

    Where carcass counts need to be made at specific short towers, we recommend using the short tower collision mortality protocols, discussed above in the methods section.  

RESEARCH RECOMMENDATIONS FOR ASSESSING RADIATION IMPACTS TO BEES

    Bees and other pollinators also deserve close scrutiny from the potential impacts of radiation, and their study should be included as part of the overall research effort suggested in this paper.  In addition to testing and validating the protocol and results from the Kimmel et al. (2006) study (see background and methods sections above), which we recommend be performed at multiple locations in the U.S., bee behavior, hive productivity, and bee survivorship need to be field-tested at both tall and short towers in the U.S.  Variations on the protocols used by Harst et al. (2006) and Kimmel et al. (2006) could easily be developed to field-test potential radiation impacts on bee navigation, flight behaviors, hive productivity, and bee survivorship around both short and tall towers.  However, any research protocol developed to assess potential insect impacts – and for that matter, impacts to birds, bats, and other wildlife, must attempt to eliminate extraneous variables that may bias study results.  These include everything from antenna placement in the Harst et al. (2006) study, to the impacts of diseases, parasites, weather and climatic events, pesticides, contaminants, and other mortality factors on insects and other wildlife.  Fine-turning a research protocol must include the combined efforts of trained entomologists, research radiation specialists, ornithologists, wildlife biologists, and biometricians.         


CONTACT:
    Albert M. Manville, II, Ph.D., Senior Wildlife Biologist, Division of Migratory Bird Management, U.S. Fish and Wildlife Service, 4401 N. Fairfax Dr. – MBSP-4107, Arlington, VA 22203.  703/358-1963; Albert_Manville@fws.gov.


LITERATURE CITED:

Arnett, E.B. 2006. A preliminary evaluation on the use of dogs to recover bat fatalities at wind energy facilities. Wildlife Society Bulletin 34(5):1440-1445.

Aronoff, A. 1949. The September migration tragedy. Linnaean News-Letter 31): 2.

Avery, M.L., P.F. Springer, and J.F. Cassel. 1978. The composition and seasonal variation of bird losses at a tall tower in southeastern North Dakota. American Birds 32(6): 1114-1121.

Avery, M.L., and R.C. Beason. 2000. Avian mortality at short (< 120 m) communication towers. Pilot Study Research Protocol to the Communication Tower Working Group. 7 pp (manuscript).

Balmori, A. 2003. The effects of microwave radiation on the wildlife.  Preliminary results, Valladolid, Spain, February.  Manuscript submitted for publication to Electromagnetic Biology and Medicine.

Balmori, A. 2004. Effects of the electromagnetic fields of phone masts on a population of White Stork (Ciconia ciconia). Valladollid, Spain, March.  Manuscript submitted for publication to Electromagnetic Biology and Medicine, with figures, diagrams and datasets not in published (2005) version.  13 pp.

Balmori, A. 2005. Possible effects of electromagnetic fields from phone masts on a population of White Stork (Ciconia ciconia).  Electromagnetic Biology and Medicine 24: 109-119.

Balmori, A., and O. Hallberg. 2007. The urban decline of the House Sparrwo (Passer domesticus): a possible link with electromagnetic radiation. Electromagnetic Biology and Medicine 26: 141-151.

Beason, R.C., and P. Semm. 2002. Responses of neurons to an amplitude modulated microwave stimulus.  Neuroscience Letters 333(2002): 175-178.

Bibby, C.J., N.D. Burgess, D.A. Hill, and S.H. Mustoe. 2000. Bird Census Techniques, 2nd edition, London. Academic Press.

Bowling, M. 2007. Where are the birds and bees? Health Action Magazine, summer: 21-22.

Cane, J.H., and V.J. Tepedino. 2001. Causes and extent of declines among native North American invertebrate pollinators: detection, evidence, and consequences. Conservation Ecology 5(1). 7 pp.

Crawford, R.L. 1971. Predation on birds killed at TV tower. Oriole 36: 33-35.

Crawford, R.L. and R.T. Engstrom. 2001. Characteristics of avian mortality at a north Florida television tower: a 29-year study. Journal Field Ornithology 72(3): 380-388.

DiCarlo, A., N. White, F. Guo, P. Garrett, and T. Litovitz. 2002. Chronic electromagnetic field exposure decreases HSP70 levels and lower cytoprotection.  Journal Cellular Biochemistry 84: 447-454.

Derby, C., W. Erickson, and M.D. Strickland. 2002. Protocol for monitoring impacts of seven un-guyed, unlit cellular telecommunication towers on migratory birds and bats within the Coconino and Prescott National Forests, Arizona. Modified Research Protocol for U.S. Forest Service. Western EcoSystems Technology, Inc., Cheyenne, WY. November 1: 9 pp.

Erickson, W.P., M.D. Strickland, G.D. Johnson, and J.W. Kern. 1999. Examples of statistical methods to assess risk of impacts to birds from windplants. Proceedings National Avian-Wind Power Planning Meeting III. National Wind Coordinating Committee, RESOLVE, Washington, DC.

Erickson, W.P., J. Jeffery, K. Kronner, and K. Bay. 2003. Stateline Wind Project Wildlife Monitoring Annual Report, Results from the Period July 2001-December 2002. Technical Report Submitted to FPL Energy, the Oregon Office of Energy, and the Stateline Technical Advisory Committee.

Evans, W., Y. Akashi, N.S. Altman, and A.M. Manville, II. 2007. Response of night-migrating songbirds in cloud to colored and flashing light.  North American Birds 60: 476-488.

Everaert, J., and D. Bauwens. 2007. A possible effect of electromagnetic radiation from mobile phone base stations on the number of breeding House Sparrows (Passer domesticus). Electromagnetic Biology and Medicine 26: 63-72.

Gehring, J., P. Kerlinger, and A.M. Manville, II. 2006. The relationship between avian collisions and communication towers and nighttime tower lighting systems and tower heights. Draft Summary Report to the Michigan State Police, Michigan Attorney General, Federal Communications Commission, and U.S. Fish and Wildlife Service. 19 pp.

Gehring, J., P. Kerlinger, and A.M. Manville, II. 2009.  Communication towers, lights, and birds:  successful methods of reducing the frequency of avian collisions.  Journal of Ecological Applications 19(2): 505-514.

Harst, W., J. Kuhn, and H. Stever. 2006. Can electromagnetic exposure cause a change in behaviour? Studying possible non-thermal influences on honey bees – an approach within the framework of educational informatics. Acta Systemica-IIAS International Journal 6(1):1-6.

Homan, H.J., G. Linz, and B.D. Peer. 2001. Dogs increase recovery of passerine carcasses in dense vegetation. Wildlife Society Bulletin 29(1): 292-296.

Kemper, C.A. 1996. A study of bird mortality at a west central Wisconsin TV tower from 1957-1995. The Passenger Pigeon 58(3): 219-235.
 
Kevan, P.G., and T.P. Phillips. 2001. The economic impacts of pollinator declines:  an approach to assessing the consequences.  Conservation Ecology 5(1). 15 pp.

Kimmel, S., J. Kuhn, W. Harst, and H. Stever. 2006. Electromagnetic radiation: influences on honeybees (Apis mellifera). Institute Environmental Sciences, Institute Science and Science Education, and Institute Educational Informatics, Univ. Koblenz-Landau/Campus Landau, Germany. 6 pp.

Levitt, B.B., and T. Morrow. 2007. Electrosmog – what price convenience?  WestView July: 7-8.

Manville, A. M., II. 2002. Protocol for monitoring the impact of cellular telecommunication towers on migratory birds within the Coconino, Prescott, and Kaibab National Forests, Arizona. Research Protocol Prepared for U.S. Forest Service Cellular Telecommunications Study. U.S. Fish and Wildlife Service, March 12: 9 pp.

Manville, A.M., II. 2005. Bird strikes and electrocutions at power lines, communication towers, and wind turbines:  state of the art and state of the science – next steps toward mitigation. Bird Conservation Implementation in the Americas:  Proceedings 3rd International Partners in Flight Conference 2002, C.J. Ralph and T.D. Rich (eds.). U.S.D.A. Forest Service General Technical Report PSW-GTR-191, Pacific Southwest Research Station, Albany, CA: 1051-1064.

Manville, A.M., II. 2007. Comments of the U.S. Fish and Wildlife Service submitted electronically to the Federal Communications Commission on 47 CFR Parts 1 and 17, WT Docket No. 03-187, FCC 06-164, Notice of Proposed Rulemaking, “Effects of Communication Towers on Migratory Birds.”  February 2.  Comments submitted to Louis Peraertz, Esq., Spectrum and Competition Policy Division, Wireless Telecommunications Bureau, Federal Communications Commission. 32 pp.

Manville, A.M., II. 2009. Towers, turbines, power lines, and buildings – steps being taken by U.S. Fish and Wildlife Service to avoid or minimize take a migratory birds at these structures.  In C.J. Ralph and T.D. Rich (editors).  Proceedings 4th International Partners in Flight Conference, February 2008, McAllen, TX (in press).

Ritz, T., P. Thalau, J.B. Phillips, R. Wiltschko, and W. Wiltschko. 2004. Resonanace effects indicate a radical-pair mechanism for avian magnetic compass. Nature 429: 177-180.

Shire, G.G., K. Brown,  and G. Winegrad. 2000. Communication towers: a deadly hazard to birds.  American Bird Conservancy Special Report. 23 pp.