3. �� Objectives of MONITORING protocol

3.1�� Pateke monitoring objectives

The monitoring protocol has a dual objective, that of quantifying demographic responses of pateke to maximum practicable management effort, and to test hypotheses of pateke response to specific management regimes.

3.2�� Protocol objective

To identify best practise pateke monitoring methods, not only at Okiwi and Mimiwhangata, but also at other sites containing pateke that will achieve the monitoring objectives above.

4.�� Monitoring parameters at managed sites

Potentially relevant parameters to measure pateke response to management include adult survival, number of nesting attempts, hatching success, brood survival, juvenile survival, recruitment, causes of mortality, abundance trends, dispersal and habitat use.� The short-term recovery strategy and work plan for pateke management and research (O�Connor 2001) identified four key parameters to measure at management sites - adult survival, juvenile survival, recruitment and abundance trend � because past work identified these as being the most sensitive population measures.�

Two other parameters, fledging success and cause of death, should also be measured.� Fledging success is important, because loss of ducklings has been extremely high at some sites in some years (Barker 1999, Brown teal workshop, Moore and Battley 2002).� In addition, the current management regimes include the control of some predator species that are likely to impact only on young ducklings. If the predator species and stage of impact are to be correctly identified, then fledging success is a desirable parameter to measure.

Recommended priority parameters for the intensively managed Okiwi and Mimiwhangata pateke populations are:

1.� �� Flock counts/abundance trends � annual counts to determine population trends in response to specific management regimes.� Increased precision of counts (particularly replication with consistent methodology) is needed and environmental variables need to be recorded and their influence reduced where possible to minimise the chances of aberrant observations.� By determining the proportion of radio-tagged adults and juveniles attending flocks (i.e. flock counts), the total population in the source area can be estimated more accurately.

2.� �� Adult survival � annual (within year) survival of a radio-tagged sample of adult pateke.� This radio-tagged sample will also contribute to determining population size (1) and finding nests and broods, and hence help determine nesting success to the late fledgling stage (3), juvenile survival (4) and cause of death of adults (5).

3.� �� Fledging success/duckling survival � percentage of hatched eggs that result in survival of ducklings to late fledgling stage.� The post-hatching inspection of nests also enables clutch size, hatching success and overall breeding success to be calculated with little additional effort.

4.� �� Juvenile survival � percent of late-stage fledglings that survive to one year.� (Note some are recruited into breeding population at Year 1, but most are believed to recruit at Year 2).� Determining survival to one and two years will be easier to achieve than determining recruitment, but the latter is desirable for population modelling purposes.

5.� �� Cause of death � determining causes of mortality of eggs, ducklings, juveniles and adults under each specific management regime. These well help determine whether the current management scenario is on the right track for population recoveries.

6. �� Dispersal patterns � determining the seasonal movements of dispersing juveniles and adults. �This will help determine the size of the management zone both spatially and temporally.

Note that over time each of these parameters is likely to be influenced increasingly by density-dependent factors.� As pateke populations increase, limiting factors such as food and nesting resources are likely to come increasingly into effect.� With flock counts, therefore, the rate of increase in size of flocks is likely to decrease as carrying capacities of the habitat are approached.� This implies that switching off the control of key predator species should not occur before key demographic patterns are fully understood - this is unlikely to be the case until several years after local populations having reached apparent �carrying capacity�.�

�� For all parameters being measured there is a need for inter-site and inter-year consistency in terms of effort and standards.� Each of the parameters to be measured is designed to answer key questions about pateke population responses to specific management regimes.� Without consistent monitoring, it will not be possible to adequately test hypotheses.� Additional monitoring is required of potential predator species and physical parameters, e.g. habitat characteristics.

5. Recommended pateke monitoring protocols for MIMIWHANGATA and OKIWI

5.1�� Flock counts

Priority	Essential
Objective	To understand the relationship between flock counts and total pateke populations (adults and juveniles) in the source area and therefore to be able to use flock counts as a means of determining trends in pateke abundance under specific research-by-management scenarios.
Monitoring target	All flocking sites at Okiwi and the Mimiwhangata area are monitored according to protocol below.� In addition, count birds at two reference sites away from Okiwi (e.g. Awana, Blackwells) on Aotea.� These latter sites are vital as controls, and the data will also serve as pre-management counts should management begin at those sites in later years.� A flocking site is a general area where more than two birds congregate during the day, particularly during the summer-autumn period.
Design	Carry out four counts per monitoring month in February-March at Okiwi (two counts 3-4 days apart during each of two high tide cycles), and in February at Mimiwhangata (single counts at approximately weekly intervals).� Historic and currently known flock sites are visited during each day of monitoring.� February-March was selected because this timing enables comparisons with census counts carried out in the past, and because, on Aotea there is evidence of this time being comparatively static for flock counts (Barker 1998). Mimiwhangata counts include all sites between Whananaki and Teal Bay.� Equivalent monitoring of at least two unmanaged reference sites on Aotea should also be undertaken on the same day (Awana and Blackwells preferred, subject to access being permitted and no concerns regarding undue disturbance, in which case use alternative site).� It appears no control sites are available in Northland. Four counts are being trialled in 2002, c.f. 1 and 2 on Great Barrier and Northland respectively in the past. Based on one year�s data, three or four counts are considered to be adequate for a high level of statistical precision (Fletcher and MacKenzie 2002).� Counts will be undertaken approximately weekly at Mimiwhangata and at as consistent a time of day for each flocking site as is possible.� At Aotea and some Northland sites away from Mimiwhangata, birds will be influenced by tides.� At Aotea the aim is for the first two counts to be undertaken 3-4 days apart in the first suitable high tide regime for the month, and the last two counts 3-4 days apart in the next suitable high tide regime. In Northland, the counts of tidal sites (Whananaki, Teal Bay, Tutaematai etc) should be timed as close as possible to that of earlier counts � this was normally, but not always, centred on mean high water spring tides (MHWS). Carry out a roll call for all radio-tagged birds at each flock site.� Determine number of radio-tagged birds away from the flock sites to establish proportion present at flock sites. Two of the counts at each of Mimiwhangata and Okiwi will coincide with and be incorporated into the �all sites counts� undertaken in February throughout Northland and February-March on Aotea. Obtain more advice on population estimation based on mark-recapture methods and adjust sampling if needed. (SCO) Search for new flock sites. Daily rainfall records are required for Okiwi and Mimiwhangata. Wind, relative humidity and evapotranspiration records are also desirable, but they may be available from nearby official stations locally.
Field methods	�Count all birds present by a consistent method for each site.� For instance, if counts have traditionally been by kayak or from a hide, keep to those methods to ensure consistency of sampling, except where the Recovery Group considers that less disturbing methods should be implemented. Document methodology for counting at each site including time and tide times and number of and role of observers.� Keep count data separate for each site. (Project leader) Visit and count all known and historic flock sites.� Be aware of potential for birds to shift sites (and for new flocks to form), so be on the look out for alternative sites during routine work.� Expand count area accordingly, but keep separate counts of any new sites that have not been visited in the past. (Project leader)� Weather � Avoid rain and/or strong winds or other events that might cause birds to be less visible. Ensure that daily rainfall records are being taken at or near Okiwi and Mimiwhangata. (Project leader) Keep disturbance of flocks to a minimum.� Any manipulation of a flock to improve accuracy of counts should be done with great care, particularly since the frequency of counts has increased.� If there is a risk of birds taking to flight, then alternative monitoring at those sites need to be implemented.� The use of hides is desirable at some sites (i.e. Burrill�s Drain and Orchard sites on Aotea, or more distant sampling to avoid close disturbance).� Similarly, telemetry can be carried out well back from the flock sites, particularly if additional observers are present. Avoid spotlighting and no catching should take place there unless it is absolutely necessary.� Where possible, keep trap-servicing sites away from immediate vicinity of flock sites. �(Project leader) Use radio-telemetry equipment for roll-call suitably removed from site to minimise period of disturbance.� Record all radio-tagged adults and juveniles as per data sheet.� Include any colour-banding observations. �(Project leader) Enter data into data sheets or note book and update database (Appendix 1). �(Project leader) Ensure staff are suitably trained and quality control occurs at start of counts annually via conservancy technical staff input or recovery group. �(Scientific coordinator, SCO)
Annual Reporting	Annual reporting should include as a minimum: Describe objectives and methods including any changes in methods. Present the mean, minimum and maximum and SE of total counts for Okiwi, Mimiwhangata and all-site areas, including flock site subtotals. Provide table of raw counts as an appendix to annual reporting. Graph mean + SE of count totals together with means (and SE if available) of previous years. In year one determine the combined % of radio-tagged adults and juveniles present at flocking sites during each count in order to estimate more accurately total �catchment� population. Analyse and graph the four total counts in relation to seasonal rainfall, e.g. total rain of past one week, one month, two months. Discuss results in relation to previous years� findings, weather patterns, etc. Discuss any problems and suggest improvements in the methods.� (All project leader with assistance from technical staff, SCO.)

��

5.2�� Adult survival

Priority	Essential
Objective	To test the hypothesis that maximum practicable management results in a high (management objective of 80%) adult survival at Mimiwhangata and Okiwi.
Performance Measure	Annual survivorship determined for 10 males and 25 females over 12 months, preferably from 1 May.
Monitoring target	n = 35 adults per annum; 10 male, 25 female.� Bias towards females assists with separate brood monitoring objective.� Note, that if there is a capacity or resourcing issue, final decisions should also consider the high priority need to locate nests for brood monitoring (objective 5.3). �
Design	Measure annual adult survival by monitoring a representative sample of radio-tagged birds.� Although this sample size is considered by Fletcher and MacKenzie (2002) to be minimal for statistical precision, inter-year replication will provide strong trends in survivorship data.� The location of captures is important � need to ensure breeding and feeding distributions are representative, therefore radio-tag birds in proportion to numbers at each of the pre-breeding feeding concentrations. It is acceptable to include the previous year�s surviving adults and juveniles in each year�s sample.� There are pros and cons of retaining the same individuals in the sample, sufficient to warrant them being included. When an adult dies, there is no need to replace it in that year�s sample, unless the sample of adult females drops to lower than 20.
Capture methods	Catch �adults� using certified pateke dog during the day in March-early May (after annual flock counts) and preferably after periods of rain when birds are in better condition. No catches should be made after mid-May when the chances of catching gravid females (which is undesirable) will increase.� If gravid females are caught in the March-mid-May sample they should be included in the sample.� If this period provides insufficient samples, the immediate post-breeding period (November-December) offers an alternative adult-catching period.� (Project leader) Stratify samples as far as possible so that the numbers of birds radio-tagged are spatially distributed in approximate proportion to the percentage of birds known or thought to be using that site, e.g. if 50% of birds are estimated to feed in a certain area then it should provide about 40-60 % of the radio-tagged sample. (Project leader) Measure effort and result in dogged samples (transect method) which will provide another monitoring tool in support of flock counts, i.e. length/area/time taken of habitat searched for x numbers of birds compared year to year. (Project leader) Certified operators attach Sirtrack long-life mortality transmitters with 15-16 month life by current backpack harness with built-in linen thread weak-link. Transmitter pulse rates should be 20/40 with an 8 hour latch time. (Project leader) Band with metal bands (females left leg, males right leg) using size S bands on most birds but size L on especially large-legged birds, and an individual colour band code. Colour bands are double wrap custom made darvic (not PVC) bands, sealed with superglue or THF (tetrahydrafuran). Banding operators require current permits. (Project leader) Measure standard body measurements � tarsus, bill, head and bill, tail, wing, weight (Appendix 3) Collect two blood feathers for subsequent DNA and paternity work � place feathers in new plastic bags without contamination, labelled, and chilled but not frozen. Photograph each bird to show general lateral plumage characteristics and label photos with band number, locality and date for future plumage analyses.� Send photos and samples to Recovery Group Leader for central storage.� (Project leader and Recovery Group Leader) After 12 months, replace transmitter during period when replacing sample of adults.� There are pros and cons for continuing with the same individuals in the sample.� On balance it is considered acceptable to include birds from the previous year because it provides an increased understanding of individuals within the population. (Project leader) Arrange training for banding operators and harness attachment (SCO, Recovery Group Leader or RGL).
Monitoring methods	Monitor at least weekly and more frequently according to seasonal need, i.e. twice-weekly for locating nests (June-September), after a predation event, or during a period of increased predator sightings.� If low mortality during summer flocking period continues, monitoring frequency in the future could be reduced to fortnightly in mid December-mid March.� Monitor birds during the day, but also include night-time monitoring in breeding season (see 5.3). See section 5.5 for processing deceased birds. (Project leader)
	Collect standard pateke data (Appendix 2). (Project leader) Train monitoring workers at the start of each monitoring year/ season and provide regular quality checks. (Technical Support Oficer or TSO, Project leader)
Annual Reporting	Describe objectives and methods. 12 month survivorship e.g. 1 May - 1 May for each of adult females, adult males, total adults as percent survival. Compare and graph mean survivorship with past years� results. Identify (tabulate/graph) timing of mortality. Identify causes of mortality � detail and summary including necropsy and other post-mortem reports for all birds appended. Discuss results in relation to weather, predator management, and other environmental factors. Discuss any problems and suggest means of improving methods. (All project leader with assistance from technical staff, SCO.)

5.3�� Hatching success and duckling survival

Priority	Essential
Objective	To determine hatching success and duckling survival (using eggs hatched and fledgling survival per sample female) of pateke in different management regimes (the first regime to be tested is �maximum practicable� effort which will be modified when monitoring targets are met and sustained).� This includes testing the hypothesis that the management target of an average of 2+ ducklings per clutch (Pateke Recovery Group minutes 2000) is being met in the current Okiwi and Mimiwhangata management regimes.
Performance measure	Survival to fledging of 30+ young from 20+ nests is established annually.
Monitoring target	n = 20+ nests from which survival to fledging of 30+ ducklings will be determined. Note: 15 nests would be the approximate statistically robust minimum.
Design	Radio-tag 25 female pateke (see 4.1). Monitor radio-tagged females twice weekly in June-September to determine nest site locations.� Hatching success � determine clutch size in immediate post-hatching period. (Project leader) Fledging success - count late stage ducklings in week before fledging (generally in weeks 5-7). Radio-tag up to two young per brood for a total of 20 juveniles (see 5.4).
Field methods	Determine whether nesting by day/evening/day fixes - constant location indicates probable nesting. �Locate nest to within a few metres, but approach to no closer than 5 m. After the nest is abandoned, count the number of eggs hatched� (from membranes visible) and failed. �In the case of an apparently deserted nest, leave the eggs in the nest and check whether they are scavenged during routine twice weekly visits, and describe and photograph sign.� Do not remove eggs, except for specific purposes approved by the Recovery Group.� This will include samples of failed eggs in the 2003 season at least � at one week past term (or immediately incubating female killed) candle eggs, then send to Massey University for necropsy following procedures in 5.5.� From twice weekly checks, estimate age of ducklings to determine approximate timing of hatching. Calculate the time (5-7 weeks after hatching) to refocus observations in the week before fledging. Count number of young in the brood (two counts or one extended count) in the week before fledging (as close to fledging as possible). Radio-tag one or two (preferably no more) young in the brood if they weigh 400+ g.� Band with metal and individual colour code and measure (5.2).� Colour codes will enable recognition and recapture if transmitter fails. (All tasks project leader.)
Monitoring	See juvenile survival monitoring below (5.4). Complete data sheets.� (Project leader.)
Annual Reporting	Describe objectives and methods. Status of individual broods, i.e. number of young surviving in each. Percentage survival of ducklings to week prior to fledging. Note details of any field observations relevant to duckling loss. Discussion of trends with weather, predator control, etc. Discussion of any problems and suggest improvements in the methods. (All project leader with assistance from technical staff, SCO.)

5.4�� Juvenile survival

Priority	Essential
Objective	To determine survival of juveniles from fledging to recruitment into the population (survival to one year and survival to first breeding).
Performance measure	20+ juveniles monitored and fates determined for 80% of them to one year of age.
Monitoring target	20 juveniles monitored at least weekly.
Design	Radio-tag a sample of 20 or more fledglings from broods of radio-tagged females, or from untagged females if insufficient radio-tagged birds.� This sample is considered minimal for statistical precision by Fletcher and MacKenzie (2002), but inter-year replication should still provide strong trends in survivorship data. Monitor juveniles
Field methods	Radio-tag and individually colour-band one or two fledglings per brood of radio-tagged females.� If total is looking like being less than 20, supplement with additional juveniles from broods of radio-tagged females, or preferably, from randomly found broods during dog searches, i.e. stratify search effort.�� Monitor weekly as for adult monitoring above.� If it is possible to monitor more frequently, especially during the September-December dispersal period, this will increase the chances of keeping track of moved birds and finding freshly dead birds for necropsy. (For dead birds see section 5.5). There is no need to replace dead birds with new birds unless the sample is drastically reduced early in the post-fledgling period, e.g. 50% loss within two months of fledging. Record data as per Appendix 2. If a radio-tagged juvenile is missing, extend coverage to outside the catchment or the area normally covered, using aircraft as necessary. Continue to monitor for missing juveniles for the life of transmitter in case the bird returns to the area.
Annual Reporting	Describe objectives and methods. Number of juveniles radio-tagged. Percentage survival to following late autumn/winter, e.g. May. Percentage survival to recruitment (breeding).� Will be a minimum figure in Year 1. Determine months and ages when losses occurred, and compare with past years, tabulate/graph. Identify causes of mortality � detail and summary as per 5.5. Discuss results in relation to previous years� findings, weather patterns, predator management, etc. Discuss any problems and suggest improvements in the methods (All project manager with assistance from technical staff, SCO)