Boxing Day Kaka Count 2011

Small Samples and Uncertainty

Towards the end of last year the GBI Trust was engaged with environmental issues with the Local Board. As a consequence a decision to go ahead with the boxing day kaka count was delayed,
so that our planning was not as thorough as in previous years. A consequence of that was that we got fewer count forms returned (23 versus 38 in 2010), and a consequence of that is that the results are much less reliable. The lesson to be learned is that to get comparable results we need comparable sample sizes. Table 1. summarises the results in the same format as last year (see Env. News # 24, Summer 2011), but with numbers in brackets added from last year for locations for which we received zero data this year! While this is a questionable procedure, we can feel fairly sure that missing counts from Kaiarara, Motairehe, Nagle Cove, Rosalie Bay, and within Glenfern Sanctuary, would have certainly added birds had they been made, and the total added (53) seems not unreasonable based on the previous counts and my observations. For example I counted ten kaka in the pohutukawa trees at Harataonga earlier in December

Table 1.
Location No Sheets Total AM Total PM Minimum Maximum
Awana 2 12 2 10 10
Port Fitzroy 1 12 - 12 12
Harataonga 1 - 2 2 13
Kaiarara 0 - - (4) (6)
Karaka Bay 1 0 0 0 5
Medlands 5 16 9 6 15
Motairehe 0 - - (7) (35)
Nagle Cove 0 - - (0) (0)
Okiwi 2 6 7 6 6
Okupu 1 - 3 3 8
Rosalie Bay 0 - - (12) (12)
Schooner Bay 2 9 7 7 9
Tryphena 3 17 17 12 12
Wairahi 3 8 23 13 13
Whangaparapara 1 6 6 6 8
Windy Hill 1 6 8 8 8
Totals 23 92 84 114 172

Table 1. Kaka count results 26/12/2011, with added data (brackets) from 2010 where no sheets were returned

Fig 1 shows that the estimated maximum total (172) is well below that estimated for 2007 (221) and 2010 (244), which is surprising because several observers noted “an increase” this year. All that can be said is that the data, such as they are, do not support the idea that numbers have increased. Comparing locations where counts were made in both years shows increases at three sites, decreases at four, and no change at the others. Thus, overall there is no evidence for an increase. I can only suggest that the birds were more conspicuous this year, because the pohutukawa trees flowered earlier and in greater abundance than usual. The previous data (Table 4, Env. News #24) recognised pohutukawa nectar as a key summer food source. Of the eleven records mentioning feeding trees this year, six referred to pohutukawa.

Estimate of Kaka Numbers

Fig. 1. Mean, maxima and minima estimates of kaka counts for GBI based on locations counted. Winter counts in white, summer counts in grey. The bars give the average, the thin lines give the maxima and minima (approximate “Error bars”).

This year we also discussed the possibility of missing kaka in our counts because they were in ‘remote’ locations on the island. As a result I did a day trip along the forest road and to Kiwiriki Bay on January 4th. I could hear kaka frequently on that trip, but my estimate of the cumulative number was only 10 -12 birds. My previous estimate of the summer total kaka population at 200-300 birds (<100 breeding pairs) may be a bit generous, perhaps 150-250 would be nearer the mark. Overall the results point out the need for a careful and standardised methodology and geographical coverage in attempting to get counts of such mobile and conspicuous birds. Thanks again to all participants, especially the dedicated folks in the Wairahi, and thanks to Emmy Pratt and Kay Stowell for helping with the organisation.

Original document: Boxing Day Kaka Count 2011