BIOTA Southern AfricaS06
The German subproject leaders are asked to incorporate the WP-reports of their African/
German R&D partners into their subproject Annual report according to the workpackages
Interim Report S06.5
Assessment of domestic grazing impacts and their sustainability in semi-arid ecosystems through the use of cheap GPS collars and G.I.S
Applicant: MsElaine Thomas/Dr Richard S. Knight (RSK)University of the Western Cape, South Africa / Promotion no.:
Short Title: BIOTA S06.5
Duration of the project: 01.01.2003 – 31.12.2006 / Reporting period: 01.01. – 31.12.2004
1List of chief scientific/technological results and other major events
Registering the Project for Ethics clearance
One problem that took some time to address was to register the project with the university’s ethics committee – who were really concerned about the safety of deploying GPS units on animals as well as the implication they could have on changing the animal’s behaviour. It was only after receiving and reviewing a full literature review and inspection of the units we intended to use was approval granted when the BIOTA contracts were signed (March 2004).
Ms Elaine Thomas was appointed to undertake this section of the research for her PhD thesis and has registered at the University of the Western Cape. She has a Zoological background with particular interest in animal behaviour.
As a start to the project and before committing BIOTA funds we did a pilot study with the first six built units from Avitrack (a subsidiary of MC Systems who supplied the BIOTA weather stations). These were purchased as an NRF funded project and deployed in Spring (September 2004). Unfortunately these had taken 18 months to actually build from the original order and even this deployment was premature. The design of the Avitrack units had the following performance specifications as compared to similar imported GPS units available at that time.
Table 1: Comparisons of various GPS collars (amount of data collected and mass of the unit)
GPS ManufacturerUnit
URL / Positions (best scenario) / Mass (grams)
AVITRACK / 60 000 / 350 (2x D Lithiums)
Environmental Systems
GPS Plus 5
/ 32 000 / 1050
Lotek
GPS 2200
/ 21000 / 900-950
Telonics
TGW-400
/ 2800 / Approx 1110
Televilt
C1200
/ 17 000 / 1250
The field trip to the Richtersveld to test these units was undertaken at a time that allowed us to join-up with Prof. Dr. Norbert Jürgens and his group. At this stage we also co-ordinated our research with a field trip by Dr Howard Hendricks from SANPARKS and who I have invited as a co-research on this project as well as a co-supervisor for Elaine Thomas. These GPS collars were deployed with extremely mixed success.
One of the major issues facing deployment of GPS collars in the Richtersveld is access to skyline to receive signals from the satellites. Normal GPS collars are programmed to actually switch on, and then search for a signal and switch themselves off until the next signal is required. This data acquisition is ideal when monitoring long-term large movement patterns of wild animals where four to six fixings per day is adequate. The problem with this regime is that while it optimizes battery longevity it generally is too infrequent to provide information useful for ecological assessment of the effects that the livestock are having on the ecosystem. Moreover, these commercial GPS units are designed so that if there is any difficulty with acquiring the necessary four satellitesfor an accurate reading they either switch off and the data is not collected or they collect data that needs to be discarded at a latter stage (two minutes is usually the maximum search permitted before the unit shuts itself down or simply records an erroneous position). A consequence of this is that at least 25% of the data cannot be used and this brings down the number of usable fixings per day.
The Avitrack GPS collars were designedwith these problems in mind and for more frequent sampling.Intervals of 10 to 15 minutes were considered the absolutely minimum period for this study. Dr Howard Hendricks has already used a commercial GPS supplied by Lotekwhich was a pre-launch unitGPS2000 (now replaced with the GPS 2200) which they claimed would last aminimum of 6 months to a year recoding four times a day – but the first unit lasted just two months. Interestingly despite the general gains in technology in the last five years GPS collars have shown little improvement with respect to operation life with the latest Lotek still only claiming one year’s operation of 6 fixings per day. Given that we already know that it is unlikely to achieve one third of this in the Richtersveld with its mountainous topography and the goat’s remarkable ability to position themselves out of reception of the GPS signals (e.g. lying in the shade of vegetation) we can estimate only 700 fixings of which only about 75% are actually usable using the latest units. This was way below the needed sampling intensity that we required for this study.
Out of the six GPS we were only able to get data downloaded from three due to a combination of hardware and software failures during the trial (including the vulnerability of the unit when a battery is swapped). A schedule of recording every fifteen minutes from 6 am to 6pm was set and the GPS were left in a sleep mode during other periods of the day (viz between recording they were not switched off to avoid cold starts which are quite draining on batteries). Unlike the early trials done with the Lotek by Dr Howard Hendricks which fixed only 76% of positions with sufficient accuracy the Avitrack units got more than 98% of scheduled GPS fixings with sufficient accuracy. Consequently we are able to secure at least fifty fixes a day as compared to 3 fixes a day that was obtained with the Lotek unit.
While this accuracy was encouraging and way ahead of other units, what was very disturbing was that it was taking on average 80 seconds to obtain a location fixing during which time a very high consumption of the batteries was experienced. Our calculations had worked on fixes not taking more than 12 secondsand the GPS could then operate for at least 6 weeks or more with the supplied 2 X D size lithium batteries but these results indicated no more than a week’s use. The hard lesson learnt was that trying to run all of the time even with fairly low battery drain of a sleep mode really flattens thelithium batteries. One of the paradoxes here is that switching off (despite the high power used during the switch-on period) allowed the batteries to “rest” and to maintain their operational life for longer periods.
It was evident that we either returned to a more conventional GPS design which would involve cold starts and obtaining much fewer fixings per day with about 75% of data being usable or we need to investigate some other way of collecting the data. This issue became further exacerbated by the cost of batteries which worked out at R680 per deployment per unit (these lithium batteries are not rechargeable). If you are using one GPS this is still feasible but with 6 GPS it is going to work out at over R4000 per week of operation – costs well exceeding our budget lines. These batteries are still far lower in cost than commercial batteries where the lowest is 180€ (Televilt) or R1 595 plus duties etc and this has half the battery capacity -1D-sized lithium.
A second major concern that we experienced with the Avitrack collars was robustness – despite being made from premium aerospace aluminium with a shrink wrapped rubber coatings they simply were not going to survive the deployment on livestock. Lotek units were certainly well designed with respect to robustness but had a mass three-times more than these units.
During a fieldtrip to Soebatsfontein we re-engineered one of the GPS and made a material covering over the GPS unit and this performed rather better but the power usage was still an issue. Observation of the behaviour of the goats wearing the collar has, however, suggested neither discomfort and rather than pushing the wearer down the pecking order (a concern raised with the ethics board) seems to have elevated its position. During the Richtersveld trial we had put the collar on Billy Goat and this might have caused more aggression by him to other animals and therefore more damage to collar through head butting.
During these trials it was evident that Elaine was struggling with the physical conditions and her health and unfortunately has been booked of sick under medical advice since October for depression and anxiety. The physician’s letter has been forwarded to the Tessa Oliver at the South Africa Biota Liaison Office, but this has not helped with the progress of the project. The problem was made more acute with the realization that a large amount of development was still necessary in order for us to deploy the Avitrack units.
During October 2004 RSK took over execution of the project and I started communicating with as many researchers in the field of GPS and telemetry as possible. Following up on the various advice received around the world I eventually came into contact with Dr Ian Hubert from the UK who sounded interested in our research problems since he was developing units for domestic stock which also required more frequent fixings and the design philosophy was more similar to what we were requiring. In ouremail discussions he made many suggestions basedon his experience for a design of a collar for our applications using dry cells (Duracells) which are easily replaced (a situation that was not the case for either the Avitrack or other commercial units). In addition he was also prepared to discount the cost provided we acknowledge the use of his products in our publications.
An extremely rigorous review of all GPS collars was undertaken including operational costs worked on our required sampling regimes and irrespective of whether the unit was designed for this operation. During this process all email correspondence was forwarded to Ingo Homburg so that BIOTA was kept in the picture. During the process I was really impressed with the honesty afforded by both Environmental Systems (Dr Uli Fielitz) and Lotek (Oana Bantus) who in the end concurred that their units were not suitable for this project
Aspects of this review are summarized below.
Table 2: Comparisons of various GPS collars based on cost of unit and operation costs during a 2-year period – costs reported in currency of the manufacturer.
GPS AnalysisBlue Sky Telemetry / Quantity / Environmental Systems / Quantity / Lotek / Quantity / Televilt / Quantity
GPS & Logger / £1,300.00 / 6 / 0.00 € / 6 / 3528.00 $ / 6 / 850€ / 6
Harness / £30.00 / 6 / 0.00 € / 6 / 0.00 $ / 0 / 6
Activity Sensors / £0.00 / 6 / 0.00 € / 0 / 0.00 $ / 6 / N/A
Temp Sensor / £0.00 / 6 / 50.00 € / 6 / N/A
Battery Packs over that supplied / £5.00 / 6 / 100.00 € / 6 / 0.00 $ / 0 / 6
UHF/VHF receiver / £500.00 / 1 / 0.00 € / 1 / N/A / N/A
UHF/VHF antennae / N/A
Interface to PC / £300.00 / 1 / 0.00 € / 1 / N/A / 0
Software / £100.00 / 1 / 0.00 € / 1 / 0
Bundle Collar and Receiver / 10,500.00€
Total Package for GPS units / £8,910.00 / 11,400.00€ / 21,168.00$ / 5,100.00 €
Total Package for GPS ZAR / R 104,413 / R 91,698 / R 108,572 / R 41,022.97
No of Positions Worst Scenario (10 min) / 4200 / 18000 / 21000 / 5000
Battery Cost per 100 positions / £0.12 / 1.39 € / 3.27 $ / 3.60 €
Table 3 Total cost to run 6 GPS Collars for 2 years including initial outlay of the GPS collars plus the cost of battery units and the number of fixings it is capable of obtaining using a 10 minute interval.
Supplier / 6 Units plus operation for 2 years (Manufacturers claim) / Replacement Battery Cost -No of fixings
Blue Sky Telemetry / R 110,523.14 / R60 – 4 200 fixings
Environmental Systems / R 140,631.04 / R2 195 – 18 000 fixings
Lotek / R 122,646.00 / R3 518 – 21 000 fixings
Televilt / R 167,856.35 / R1 595 – 5 000 fixings
Per unit the Televilt units were far the cheapest but since they could store few fixings and each battery was expensive for its capacity its operation cost made it the most expensive in the long term. Essentially Televilt sell their units at cost or below cost and make profits on battery packs. Although Lotek offered a lot or reprogramming of their units to cater for what we wanted no one was sure how long their batteries would last in the field. Environmental Systems forwarded all of their software including a program to calculate expected battery endurance and actually would have been the second choice since they were going to donate a second-hand receiving station to the project. Only Blue Sky Telemetry units designed by Ian Hubert was within our original budgetwhen including operating costssince the batteries are normal Duracell Ultras and easily obtained and replaced in the field (which they will have to be since they will obtain only about 4200 fixings). This we will actually require working more closely with the herders and to train them to replace the batteries so that more continuous data is obtained.
During November an order was generated for 6 Blue Sky Telemetry custom-designed GPS collars which also have activity and temperature sensors1. Further we can obtain information from each GPS unit remotely through radio telemetryand since each unit has its own radio frequency we can deploy more than one GPS collar in a single herd of goats and monitor individually and remotely their positions and activities.
1 Additional note the GPS unitsarrived and where cleared through customs at the end of March 2005
2Comparison of project status with original work, time and cost planning (or with planning amended with the consent of UHH, respectively German R&D partner institution)
The work is behind due to the need for sourcing another supplier for the GPS collars – this information was passed onto UHH (Dr Ingo Homburg) and his approval obtained for the change in supplier of the GPS unit and to deploy fewer units but on a full time basis. Originally we were going to purchase 12 units and have 6 units running at any one time and these would be replaced with the other six units since the batteries have to be replaced at the factory. With these one we have ordered the same 6 units are permanently deployed since the batteries can be replaced in the field.
While the original intention was to develop loggers that are smaller and simpler using a new approach and directly related to semi-domesticated rather than wild-animals this has had to be dropped since Avitrack has now no real interest in project due to the hugely improved exchange rate which has made their local manufacture uneconomic – especially with the introduction of Televilt’s 850€ budget unit (Table 3). Effectively Blue Sky Telemetry has stepped in with extensive customizing of their units for our research and almost certainly more environmental data can now be obtained together with the positioning data. The only negative is the units are relatively heavy at 1050 grams.
The second reason for not being online with the original work is that the principle researcher (Elaine Thomas) became very ill from late October 2004. Unfortunately the prognosis for her improvement is not encouraging in the short-term2.
2Additional note a meeting was held April 2005 in which Elaine Thomas agreed to resigned from the BIOTA project and has returned BIOTA purchased equipment (Mecer Laptop).
3Have the prospects of reaching the project targets within the stated cost period changed compared with the original proposal (reasons)?
We believe we can still get the information required as per original workplan and actually get better data since activities and temperature can also be recorded3
3Additional note Dr Howard Hendricks has agreed to conduct the research in lieu of Ms Elaine Thomas and to become a research associate of the BCB Department at UWC. This will ensure that the project will reach original targets providing the BIOTA agrees to this arrangement.
4Have any third parties announced R&D results of relevance for implementation of the project?
No
5Are any changes to the targets necessary or will any become necessary?
No
6Have licenses or patents been issued?
No
7Is there any economic use of the outcome (even after project end). If yes, when will it occur?
Cannot be foreseen at this stage.
8What might be the potential success (scientific / technical) after project end, how can results be used (for public issues, data bases, networks)? Please consider co-operation with third parties (institutions, companies, networks, research facilities).
This project will contribute to the management of the RichtersveldNational Park through getting the most detailed data on livestock activity patterns that current technology can provide.
Since the deployment of the GPS collars on livestock requires inputs from the herders themselves it will reinforce the concept of partnership and shared responsibilities for park management.4
The research will also be prepared for publication in Scientific Journals.The data will also be made available via the Meta Database of BIOTA Southern Africa ( and obtained free of charge by bona fide scientists that signed the Data Sharing Protocol of BIOTA AFRICA.