tfic_edyvane_giant_kelp_comment.htm  

Introduction
Published scientific reports should provide validation
In contrast to today's world of uncertainty, many scientific techniques and methods are refined and validated as being able to give a high degree of certainty in the accuracy of the data used and the veracity of the analysis of that data. In published scientific papers, statements and claims made from project data analysis and other methods of assessment should be backed up by references to other published papers or otherwise validated sources. This type of work should (without exception) be unbiased, and reflect a true and open-minded interpretation of the results found through the various methods of assessment.

It is important to look at the Giant Kelp Report's claims of extensive kelp bed 'loss'. TFIC contends that it is not necessarily 'decline' or 'loss' at all, being mindful that commercial fishermen are on record as stating that this is a species that can demonstrate substantial fluctuations in bed area, even to the point of being non-existent for a decade and then returning with vigorous growth. The data used in the report tends to support this (rather than some of the conclusions drawn).

The report appears to partially recognise this in 3.1, but some of the conclusions drawn from data analysis within the report are certainly not reflective of it.

• 3.1 (Pg 73; Para 7): However, any detailed spatial-temporal (GIS-based) analysis of the historical distribution and current status of Macrocystis pyrifera in Tasmania, must the (sic) recognise the dynamic and gross inter-annual fluctuations of Macrocystis, (which are linked closely to environmental factors and resource availability), and also, should assess and integrate the wide range of quantitative and qualitative, and often disparate, data sources (and their limitations).
The public launch of the Edyvane report
The public launch of this report was contentious amidst claims of it being 'leaked' due to its confronting nature. Media claimed that the Giant Kelp report stated that 'overfishing' was the reason for the so-called massive decline in Giant Kelp beds, also claiming that Minister Bryan Green had seen the report weeks previously and was 'sitting on it'. At this point, even some members of the Steering Committee had not seen the Final Report that was finally produced several months after it was due. When asked for comment by the media, this mishandling of the launch left the Tasmanian Fishing Industry Council (TFIC) at a distinct disadvantage. Also, upon query, the Minister indicated that he only had the report sent to him around the same time as the media event.

'Overfishing'
Media events since the release of the Edyvance report have declared that commercial 'overfishing' is a cause of kelp decline. Even scientists associated with the production of this report and others who assess these same fisheries strongly disagree with this claim. Oddly enough, the report alludes to it in very soft language, making no claim and offering no scientific evidence. Yet the media still gets told that overfishing is a cause. This is something that the media should check before they publish. This is a claim easily proved to have little basis in reality.

The value of the Edyvane Report
The value of this report is extremely questionable. The data that was used was gained by various methods, one method often contradicting another with substantially different values. The data had severe constraints and limitations but was as good as could be expected when using archival information from varying sources and scales. However, under even a cursory scrutiny, there appears to be selective use of the data and astonishing conclusions drawn from this unique method of data interpretation. As well as this, the Giant Kelp species Angustifolia was not included in this 'desktop' assessment of Macrocycstis pyrifera and has since been genetically proven to be the same species. Angustifolia has extensive beds in areas around Tasmania. With this exclusion alone, the report is flawed.

This TFIC assessment of the report lets the words and data within the report speak for themselves, but even a most charitable description has to refrain from describing this as 'good unbiased science'.

Note:
To aid the readers understanding of this assessment of the report...

• This assessment of the Edyvane 2003 (Final Report) on Giant Kelp uses selected verbatim quotes and citations directly from the report.
• A 'dot-point' and reference to where they can be located within the report always precede these quotes.
• Any typographical or other errors detected within the original text of the report remain unchanged and are noted with "(sic)".
• Any text omitted within a quote is indicated with "…".
• RM (to assist in emphasis of points raised) has added all underlining.
• All comment by RM is in Italics.



---

Decline in Giant Kelp beds (as described in the report):
Causes


• 1.4.1 (Pg 28; Para 1): Natural oceanographic and physical influences such as El Niño and La Niña episodes, storms, rainfall… and human associated influences such as discharged wastewaters.

  
  (Para 2) In NZ, M.pyrifera does not persist in areas where the maximum temperatures exceed 18 - 19ºC for several days, and where the warmest monthly isothere does not exceed 16 - 17ºC.
  (Para 3) In southern California, declines in water quality related to turbidity from coastal development, ocean discharges, and non-point source runoff have caused significant reductions in the areal extent…
  .

• 1.4.3 (Pg 30; Para 1): Episodes or seasons of storm waves, which cause structural failure of M.pyrifera plants are probably the most important source of mortality for adult Macrocystis in California…
  
  .
• 1.4.4 (Pg 31; Para 1): In California, grazing by sea urchins is another major source of kelp mortality… When urchins reach high population numbers they are known to completely remove Macrocystis from an area, resulting in an 'urchin barren'.
  
  .
• 1.4.6 (Pg 33; Para 8) and 5.2.4 (Pg 141): Japanese kelp it (sic) thought to pose a major threat to endemic algal communities, such as M.pyrifera because of its rapid growth and capacity to exclude native species by overgrowing… RM Comment: Perversely, the introduced Japanese kelp is considered to be a threat in these paragraphs, yet 5.2.4, (Para 3) says: "Importantly…Macrocystis can establish and recover within Undaria forests, if there is no sediment or disturbance."
  
  .
• 2.10.3 (Pg 61; Para 3): As such, during 1970-71, kelp production (ie. standing stock biomass), decreased significantly (ie. from 8,120 in 1970 to 2,825 wet tons in 1971), but rates of kelp harvesting (ie. harvest as % production) increased significantly (ie. 2-6 times) over this time period.
  
  .
• (Pg 63; Para 4 to 6): In March 1972 a major kelp loss event occurred along the southern and eastern coasts of Tasmania (see Figure 2.10). Field survey notes on the 29 March, 11 April, and 17 April, 1972, reveal that many kelp beds on the east coast of Tasmania, ie. "northern beds" (Eddystone Point to Friendly Beaches), Bruny Island, and the "southern kelp beds" (Dover to Recherche Bay) were completely decimated by heavy seas caused by a tropical cyclone in the Tasman Sea on the 25th March. Only 700 tons available from Tasman Peninsula to Eddystone Point." (sic) (Bond 1972):
  
  "It is apparent from this survey that all the northern beds have suffered heavily as a result of the heavy seas caused by the cyclonic depression of the 25th March. In all areas surveyed many hundreds of tons of cast weed are lying on the beaches, in places up to four feet deep. The one consoling feature of this situation is that much of the cast weed appears to be old growth and hopefully much of the young regrowth has not been damaged." (David Bond, 17^th April 1972).
  .
  Recovery of the kelp beds following the major storm event of 25th March 1972 was very slow. In May 1972, Bruny Island kelp beds did not show any sign of recovery (or "regrowth) at all, with minimal "regrowth" in Port Arthur and Fortescue Bay. This is confirmed in the field survey data (see Table 2.12). In the 8th June 1972 survey of the northern kelp beds, no recovery of beds was observed:
  
  "The fact that concerns me most is that in all areas surveyed, the remaining weed is all old growth and there does not appear to be any young regrowth at all." (Bond, June 9th 1972).
  .
  By 15th June 1972, "regrowth" in Port Arthur had improved greatly, with a fourfold increase in available kelp since the previous survey of May 11 1972 (ie. a 5 week period), with the beds "ready for harvesting". By July 19th 1972, there were signs of recovery of the Bruny Island kelp beds, with Adventure Bay particularly, showing good potential for harvesting (Bond 1972).
  .
• 3.1 (Pg 73): In 1987 the quantities of Macrocystis pyrifera decreased dramatically, again coinciding with increasing annual mean seawater temperatures (SeaCare 1999). In 1988, the total amount of M.pyrifera on the coast of Tasmania was estimated to be less than 50 ha (SeaCare 1999). By 1996, the levels of the kelp had slowly increased to approximately a quarter of levels surveyed in 1986 (SeaCare 1999) (see Table 3.1).
  
  .
  In 1999, a survey by SeaCare (Sanderson 1999) observed significant recovery of kelp beds along the east and south coast of Tasmania, with stocks of Macrocystis as high (RM: 'higher'?) in surveyed areas than at (RM: 'as'?) any other time since 1983. However, little evidence of kelp regrowth was noted on the central east coast (ie. Marion Bay to Binalong Bay region), despite previously (sic) records of very large beds (Cribb 1954, Button 1961 cited Sanderson 1987, Sanderson 1987). Notwithstanding the recent recovery of Macrocystis beds observed in 1999 via aerial survey, a number of possible causes or factors of kelp decline have been identified by researchers in Tasmania, including:
  • large scale oceanographic changes, specifically the increased penetration and influence of the warm-water, low-nutrient Eastern Australian Current southwards along the east coast of Tasmania, which has resulted in a 1.5ºC rise in sea temperature since 1940 (Edgar 1999, Crawford et al.2000);
  • large scale ecosystem changes along the east coast (as a result of the above oceanographic changes), particularly the increased abundance, since the 1960s, of the black sea urchin (ie. Centrostephanus rodgersii), which are known to graze on large kelps (Edgar 1999);
  • the effects of marine pollution, particularly in the Derwent estuary and D’Entrecasteaux Channel (SeaCare 1999);
  • and/or the introduction of the Japanese Kelp (Undaria pinnatifida) on the east coast of Tasmania, which has colonised many areas formerly occupied by M.pyrifera (Sanderson 1987). (RM: This is at odds with 5.2.4)
  .
  Other potential causes of kelp loss, particularly on the central east coast of Tasmania, include, the commercial harvesting of kelp in the 1970s, coastal runoff, scallop dredging in the 1950s, and ecosystem changes due to fishing (Sanderson 1999).
  .

---


Limitations and constraints in the data used in this report:

RM Comment: It is important to note that a significant proportion of this report is taken up describing the multitude of problems, limitations and constraints inherent within the various methods and/or data sets used to assess Giant Kelp beds. As described above, recognised within the report are the major and extreme fluctuations/variability in both growth and area of Giant Kelp beds.

Ground-truthing, (i.e., divers or vessel towed cameras etc) physically identifying or confirming the presence (or absence) of Giant Kelp beds appears to have been able to graphically prove inaccuracies or limitations in all of the other methods of assessment. For example:

• 3.3.2 (Pg 91; Para 2): Recent field ground-truthing (2001) confirmed beds at Pierson Point, Ninepin Point and Garden Island Point, and also identified several previously unrecorded beds at Woodbridge, Pensioners Bay, Arch Rock, Huon Island, Verona Sands, Garden Island, Garden Island Bay, Randalls Bay, Cray Point, Eggs and Bacon Reef, Cygnet Point (Eggs and Bacon Bay).
RM Comment: (This is not an isolated incident. See also Pg 90; Para 5, quoted below). These and other considerable disparities and problems encountered within the assessment process must (by their very nature) indicate that the conclusions drawn are both nebulous and questionable. Another example of limitations and problems with the data assessment…
• 3.2 (Pg 74; Para 2): Estimates of Kelp Loss: Estimates of kelp loss were determined primarily using data from aerial photographs and Landsat imagery. While data from historical navigational charts was often highly accurate in terms of defining the linear coastal extent of kelp beds, it was inaccurate in estimating the seaward or offshore boundary of beds. As such, the seaward boundary of polygons were often located over non-reefal habitat (sand), ie. non-kelp habitat. Similarly, data from aerial surveys was accurate for small, well-defined kelp beds in bays and islands, but was inaccurate and inconsistent for open areas of coast and long stretches of coast…
RM Comment: This last statement seems inconsistent with claims made in 2.10.4 (Aerial Survey Data); and 3.3.2 (Pg 97; Para 2) which states that "Significantly, ground-truthing (2001) has confirmed beds at all the beds identified by recent Landsat TM and aerial survey." (Obviously specific to that area)
.
• 3.3.1 (Pg 75): The distribution and abundance of Macrocystis pyrifera along the east and south coast of Tasmania has fluctuated considerably over the past 5 decades. Estimates of total areal extent have varied from 11,986 ha (recorded in the early 1950s), to a minimum of <50 ha recorded by aerial survey in 1988 (SeaCare 1999)… The most recent estimates (ie. 1998- 2000) indicate a total area of between 558 ha – 1,311 ha, depending on the mapping methodology.
  
  .
  However, some of the kelp areal estimates are questionable. For instance, the estimates of CSIRO in the early 1950s (Cribb 1954) are probably over-estimates, as proposed by Sanderson (1987) (see Section 2.10.2). This is supported in this study. In addition, the areal estimates by aerial surveys (Sanderson 1987, Sanderson 1999) are probably accurate for kelp beds in bays and around islands, but over-estimates for large offshore beds, and narrow, long inshore reefs...
  .
• 3.3.2 (Pg 90; Para 5): Early navigational charts of Norfolk and Frederick Henry Bays (1893), identified beds at Carlton Bluff, Isle of Caves, Fulham Island, Betsey Island (north coast) and Black Jack Rocks. Recent field ground-truthing (2001) confirmed these beds as well as identifying beds at Renard Point, Spectacle Head, Spectacle Island, Whale Rock, Cape Direction and the Iron Pot. None of these beds were detected using recent Landsat TM (Sep_1999) imagery. Similarly, recent aerial survey data (Aug_1999) only identified a single kelp bed at Iron Pot (10,385 m²). In addition, no kelp beds were recorded at all in this region in the 1986 aerial survey.
  
  .
• 2.1 (Page 39; Para 1): Giant Kelp forms a buoyant canopy at the sea surface which can become extremely dense when mature, making it well suited to aerial inventory techniques.
  
  .
  (Para 3): Kelp growth … is strongly coupled to environmental factors, particularly water temperature and nutrient levels. …while growth generally peaks in late winter - early spring, maximum canopy production cannot be generally predicted, due to the strong coupling with environmental factors..
  .
  2.1 (Pg 40; Para 2): Canopy cover area or abundance however, is only a measure of the relative abundance of kelp beds and does not provide an overall estimate of the biomass or density of the submerged kelp beds.
  .
• 2.4.2 (Pg 45; Para 4):(RM: This refers to mapping that was carried out in 1954.) Open water inside the outlying buoys was not excluded from the areal estimates, and hence, most of the canopy area estimates are significantly over-estimated.

---


2.6: Aerial Photographic Mapping
• 2.6.2 (Pg 47): Methodological Limitations
The examination of the aerial photographs presents several problems of both, interpretation and scale:
• The scale of the photographs ranged from 1:5,000 to 1:42,000. This meant that some kelp beds apparent in the smaller scale photos were undetectable in the larger scale photos…
• In some cases the detection of Macrocystis was impossible due to obscuring cloud (although this was infrequent) and more frequently to reflection of sunlight from the water surface or through wave action…
• The photographs suffered a loss of resolutions (sic) once they were scanned…
• It was not possible to determine if any kelp was beneath the surface.
• Thus, the maps generated represent only the large beds.



---

2.7: Landsat Mapping

• 2.7.2: (Pg 50) Accuracy Assessment: In some regions, Landsat interpretation revealed a greatly reduced extent of kelp beds than (RM: 'when compared to'?) aerial photography. However it is difficult to discern how much of this difference is related to the difference in acquisition times, how much is due to the failure of Landsat to detect kelp… For instance, satellite mapping in Fortescue Bay indicates a greatly reduced extent than (RM: 'when compared to'?) the aerial photography acquired 16 months later… Similarly, in Southport, aerial photography acquired 16 months after the Landsat image, reveals a large bed not detected by Landsat, but (RM: 'and'?) smaller beds in some areas.
  Mapping Limitations:
• In some circumstances (ie. shallow waters), it may be difficult to distinguish Macrocystis pyrifera from M.angustifolia and also, other kelps with surface canopies…
  
  .
• Also, without further work to determine the relationship between what is visible in a Landsat image and the actual area of kelp canopy (either surface or immediately subsurface) it will not be possible to give highly accurate estimates of canopy area.
  
  .
• Low sun angles (ie. shading by cliffs) may confuse the signal for Macrocystis.
  
  .
• As the Landsat images are already acquired (approximately ever (sic) 16 days), the time of year cannot be chosen. This means that some images may not coincide with maximum canopy area.
  
  .
• There was a considerable decline of Macrocystis canopies and beds around 1987/88, which may reduce the accuracy of subsequent classifications.
  
  .
• 2.10.1 (Pg 53): Despite the acknowledged mapping limitations, aerial photography captured an accurate picture of the historical distribution of Giant Kelp, Macrocystis pyrifera, along the east and south coast of Tasmania. (RM: A claim such as this would need to be have good quality validation when one considers the significant limitations and problems described in this report for this method of data collection and analysis. This is an unquantified claim needing support. Where is it?)
  
  .
• (Pg 54; Para 2): In contrast, the recent estimates from aerial surveys indicate a total area of 2,562 ha, ie. 333% greater than the combined estimate from Landsat TM and aerial photography. (RM: This method is described as being validated by ground-truthing of survey data collected for the aerial survey for Southport to Recherche Bay…Page 97; Para 2).
  
  .
• (Pg 54; Para 3): When the number of kelp beds (ie. polygons) are examined, aerial photographic mapping identified fewer beds (68), but with larger canopy areas (mean = 81,996 m²), while Landsat detected more beds (239), but of smaller areal extent (mean = 14,697 m²) ...
  
  .
• 2.10.2 (Pg 55; Para 7): The area and productivity estimates by Cribb (1954) of Giant Kelp, Macrocystis pyrifera, for the south and east coasts of Tasmania, were vast overestimates. Cribb (1954) estimated a total area of approximately 11,986 ha and a standing crop of approximately 118,000 wet tons at one harvest per year (see Tables 2.2 and 2.7). These estimates resulted in the commercial interest in the harvesting of Macrocystis pyrifera for alginates. However, in the early 1960s, when Alginates Australia conducted a survey prior to harvest, they recorded approximately 1,215 ha and a possible annual tonnage of 11,000 wet tons.
  
  .
  (Pg 58; Para's 1 to 4): Kelp mapping by Cribb (1954) was undertaken using buoys to mark the outer edge of the kelp bed, and the area measured using a range finder. Open water inside the outlying buoys was not excluded from the areal estimates, and hence, most of the canopy area estimates are significantly over-estimated (Mick Olsen, personal communication). Alginates Australia (ie. Mick Olsen) undertook an aerial kelp survey on 6 December 1960 and 21 July 1961, and estimated the discrepancy was about the order of 10 for the amount of kelp for areas south of Hobart…
  .
  The veracity and authenticity of the original estimates of quantities of Macrocystis pyrifera made by Cribb (1954) were the subject of considerable, and ongoing, formal correspondence between Alginates Australia and the Fisheries Division (Department of Agriculture), the former State Government department responsible for managing the kelp harvesting industry.
  .
  Sanderson (1987) in a latter aerial survey of Macrocystis pyrifera stocks on the east coast, also doubted the estimates by Cribb (1954), and recognised that several of the larger beds recorded by Cribb (1954) were located within large sandy bottomed bays, such as Oyster Bay (Maria Island). Other areas however, had little or none where there appears to have been large quantities in the past (The Gardens, Okehampton Bay, Stapleton Point) (Sanderson 1987).
  .
  When the kelp maps of Cribb (1954) are closely examined in relation to the current known extent of reef habitat (based on recent 1:25,000 and 1:100,000 habitat mapping), there are large discrepancies. Some of the largest discrepancies include Oyster Bay, Hermitage Point, Middle Bluff and Carrickfergus Bay. While Cribb (1954) generally accurately identified the location of beds, the principal problem appears to have been the delineation of the seaward boundary of the kelp beds.

---

2.10.4 Aerial Survey Data (Page 68)

• Aerial surveys conducted by Sanderson (1987, 1999), were generally very accurate in locating kelp beds along the east and south coast of Tasmania. Surveys were particularly accurate in identifying small beds. For instance, beds as small as 21 m² (off Maria Island) were identified in 1999. In addition, aerial survey data has provided kelp records in locations where recent aerial photography and/or Landsat TM data has not been available. This is particularly the case for the central east coast (such as St Helens to Bicheno, Freycinet Peninsula – Schouten Island region), where recent aerial photography was not available. As such, aerial surey (sic) data has resulted in many previously unrecorded beds. In addition, aerial survey data has provided invaluable corroborative evidence for kelp records based on aerial photographic and Landsat TM data.
.
RM: It is interesting to note that this report has a very selective attitude towards aerial survey data.
.
However, the aerial survey data is limited by the dataset, which includes only 2 sampling periods (June 1986 and August 1999). Further, the area north of Friendly Beaches was not surveyed in 1986…
.
One of the most significant problems with the aerial survey data however, is the level of accuracy when compared to other kelp mapping methods (see Section 2.10.5). Data was accurate for small, well-defined kelp beds in bays and around islands, but was inaccurate and inconsistent for open areas of coast and long stretches of coast.
.
(Para 4): In addition there are significant discrepancies in the published data (Sanderson 1987) and the GIS data for the June 1986 aerial survey.
.
• 2.10.5 (Pg 69; Para's 4 to 6): When estimates from aerial photography and aerial survey data are compared for individual kelp beds, it is clear that aerial survey data does not correlate well with data from aerial photography (Table 2.15, Appendix 10). While the aerial survey data generally overestimated the aerial photography estimates (by an average of 3.52), the variance was high (SE of 10.01), with the aerial survey data both, significantly underestimating (down to 0.001) and overestimating (up to 67.62) the estimates obtained from aerial photography.
  
  .
  The largest discrepancies in area estimates for kelp beds occur when Landsat TM (1999) and aerial survey data (1999) are compared (Table 2.15, Appendix 10). Aerial survey data generally overestimated the kelp area estimates from Landsat (by an average of 60), with a very high variance (300), consistently.
  .
  As such, data from aerial surveys was accurate for small, well-defined kelp beds in bays and islands, but was inaccurate and inconsistent for open areas of coast and long stretches of coast...
  .
RM Comment: Again, this last statement seems inconsistent with claims made in 2.10.4 (Aerial Survey Data); and 3.3.2 (Pg 97; Para 2) which states that "Significantly, ground-truthing (2001) has confirmed beds at all the beds identified by recent Landsat TM and aerial survey."
.
2.10.6 (Pg 70) Mapping Limitations
• Timing of Canopy Formation: The techniques developed to identify kelp in the 28/09/1999 Landsat image for south east Tasmania could not be successfully applied to the remaining six images and epochs. The remaining images gave no indications of discernable areas of M.pyrifera. The failure of the developed methodology to detect the presence of M. pyrifera in the other images is most like (sic) due to the time of image acquisition. All of the Landsat images other than the 28/09/1999 image (ie. south east Tasmania) were acquired in summer, with two acquired in late summer, April and March. Of the summer images, only the earliest (08/01/1988), showed any indication of kelp related changes in reflectance. To this end, recent Landsat TM data was not available for the north east coast of Tasmania.
  
  .
  As such, Landsat TM was most effective in identifying kelp, when the image was acquired in spring (September_1999) or early summer. Reduced canopy area decreases the reflectance in the near infrared. As a result, the ratio of near infrared / red reflectance is not sufficiently different from open ocean to enable mapping of the kelp canopy.
  .
  Low Sun Angles: Low sun angles resulting in shading of waters below high cliffs may have contributed to the occurrences of kelp like (sic) signatures close to the coastline. However, the extra 25m buffer around the coast would have reduced some of this effect. Importantly though, the low sun angle of the September image (solar elevation of 39º) did not appear to impair the capacity of Landsat TM to detect kelp.
• (Pg 71) Kelp Area and Density: Extensive fieldwork would be required to determine any relation ship (sic) between kelp density and values of the ratio of bands 4/3. – ie. if the potential measures of Kelp density (TM band 4 / TM band 3) provide valid indications of actual canopy density. Further fieldwork would also be required to determine how the bed area derived from the Landsat image relates to the actual size of the M. pyrifera bed.
.
RM: i.e., ground-truthing due to the extreme uncertainty about the data.


---

Some examples of questionable data interpretation…

• (Pg 78) Binalong Bay and the Gardens
  
  .
  Binalong Bay and the Gardens was the subject of a detailed temporal analysis, based on data from aerial photography (1964, 1977, 1984, 1993, 1998 and 1999), and corroborated by data from aerial surveys (1999) and kelp harvesting data from Alginates Australia (1973).
  .


Table 3.4 Distribution of Giant Kelp (Macrocystis pyrifera) in the Binalong Bay and Gardens region (1964 -1998).

.
• Kelp beds in the Binalong region have declined in the period 1964-1998, from a total area of 1,477,202 to 760,448 m² (ie. 48.4%) (see Figure 3.1, Table 3.4). This is largely the result of contraction of existing beds. As such, the average size of the beds in the region have declined from 147,720 m² in 1964, to 69,132 m² in 1998.
.
RM Comment: It is astonishing to read a claim such as this from the data presented. If the data in the table above is to be believed (doubtful if all of the problems with the data listed above are also to be believed), in 1964 an aerial photo was interpreted as showing Giant Kelp covering 1,477,202 m².

In the ensuing 13 years to 1977, data obtained from Alginates Australia (involved in harvesting Giant Kelp) demonstrated the fluctuation and variability expected in a species such as this. It is likely that this data is not so accurate, but it shows fluctuations.

At the end of 1977, another aerial photograph (not a survey) showed an almost identical area covered by Giant Kelp as the 1964 photograph. Subsequent years have shown typical fluctuations, as is to be expected. So how can the claim possibly be made that the beds have declined by 48.4%? It is a far more realistic interpretation of the data to conclude that this is a natural fluctuation as is expected from this species.

The comparison between 1964 and 2000 may show reduction in the size of the beds, but to claim it as a decline is totally incorrect. It clearly demonstrates fluctuations, something that is quite normal for this species.
.
.
(Pg 80) Freycinet Peninsular and Schouten Island.


.

• (Pg 81): Kelp beds in the Freycinet region have declined considerably over the period 1948-1991, from a total area of 1,657,069 m² to 340,241 m² (ie. 79.5%) (see Table 3.5). This is largely the result of both, loss of kelp beds (ie. Wineglass Bay, Bluestone Bay, Little Bluestone Bay, Carp Bay, Sleepy Bay), and the contraction of existing beds (ie. Slaughterhouse Bay, Friendly Point, Gleaner Reef). As such, the average size of the beds in the region declined from 103,567 m² in 1948, to 68,048 m² in 1991. Recent aerial survey data (1999) supports these findings, with large scale loss of kelp beds along most of the coast, apart from 2 large beds at the Friendly Beaches, and 2 small beds persisting in sheltered areas.
  
  .
RM Comment: The conclusions drawn from this data set cannot be considered to be realistic. In 1948 there was photographic evidence interpreted as representing Giant Kelp beds covering 1,657,069 m². This is about where the reality starts and stops in this highly questionable interpretation. Twenty years later, 1968, there was evidence of a twofold increase, to 3,037,948 m². Despite the inclusion of pessimistic data from Alginates Australia that demonstrates a significantly smaller area covered by Giant Kelp, by 1978 the beds had expanded even more to cover 4,257,738 m²(aerial photograph, not aerial survey). That is an astonishing recovery and growth rate for a five-year period.

Why then, is the report concentrating on using data from Nov 1991 to claim a 79.5% decline, when just under eight years later there is 1,139,227 m² of Giant Kelp beds evident again?

Surely this proves the extremely variable nature of this species. There is no indication of a 'loss'. It clearly demonstrates natural fluctuations, not decline. And of course, if the 1999 data were used, the 'decline' would not have been nearly as spectacular!
.

3.3.2: North Bruny Island
.


.

.
(Pg 92; Para 2): Kelp Beds on the coast of North Bruny Island have declined over the period 1947 - 1999, from the total area of 877,961 m² to 649,785 m² (25.9%)… This is the result of loss of beds and to a lesser extent, contraction of existing beds. Large beds have contracted at Variety Bay, while recent Landsat TM identified small, persisting beds at Kellys Point, Patricks Blight, Bleaches Bluff and Yellow Bluff. However while these latter beds are identified with early aerial photography (1947) and historic charts (1860), and aerial survey data (1999), they were not detected with recent aerial photography, nor have they been verified with recent field ground-truthing (2001). Ground-truthing (2001) has only confirmed beds at Kellys Point (Cape De La Sortie), and Lookout Bay to Variety Bay.
.
RM Comment: Table 3.15 shows an interpretation of a 1947 aerial photograph that represents kelp beds covering 877,961 m². The report claims a loss of kelp beds of 25.9%, basing the assessment on data from 1947 and 1999 alone.

What about the fact that other aerial photographs from 1980 to 1987 show fluctuations as low as 338,025 m² (1985) followed by a threefold increase in the next two years to 974,444 m² (1987) that is greater than the original 1947 assessment? Why is there such an obvious need here for the selective use of data?

.
3.3.2 (Pg 93) South Bruny Island
.


.

• (Pg 93; para 5); The kelp beds of the east coast of South Bruny Island (in contrast to North Bruny), have undergone considerable decline over the period 1947-1999, from 1,978,348 m² to 176,487 m² 91.1%) (see Table 3.16 and Figure 3.6). Major losses have occurred at Adventure Bay and also, south of Fluted Cape.
  .
• (Pg 94; Para 2): Recent field ground-truthing (2001) and Landsat TM (Sep_1999) has also identified significant populations of Macrocystis on the west coast of South Bruny Island, particularly in Great Taylor Bay and in the Big Reef to Simpson Point region (including Satellite Island). While the taxonomic identify (sic) of these plants has not been confirmed, they are most likely Macrocystis pyrifera. As such, specimens of kelp collected by Scoresby Shepherd in Great Taylor Bay in 1970 (at 2-7 m depth), have been confirmed as Macrocystis pyrifera (see Womersley 1987).
.
RM Comment: To claim a 91.1% 'loss' over this time frame is unrealistic. The data clearly shows fluctuations in bed size, peaking at 1,357,850 m² (survey) in 1986, reducing dramatically in size over the next ten years to 185,173 m².

Three years later in 1999 the data shows an equally dramatic increase in bed size to 1,316,261 m² (survey), decreasing again substantially in the next few months.

Is this a result of one or a combination of the problems inherent in the assessment methods? Or is this simply the natural variability of this species, perhaps dictated by natural ocean processes such as warmer waters or storm events?

To make the assumption (without accurate validation) that these populations of Macrocystis are M.pyrifera raises another question. Without species identification in all the data used, how many beds are another species of kelp such as M.angustifolia? Or are they the same alga, as questioned by Craig Sanderson? (Personal communication Dec 2003).

.
3.3.2 (Pg 96) Southport to Recherche Bay
.


.

• (Pg 97; para 3): Kelp beds in the Southport to Recherche Bay region declined over the period 1946 -1999, from a total area of 1,507,636 m² to 236,250 m² (95.6%) (see Table 3.18 and Figure 3.7). However this is likely to be an over estimate of the loss, as the recent aerial photography was taken in summer (Feb_2001), a time of minimum canopy formation. When the recent Landsat data (Sep_1999) is examined (albeit with a lower scale of resolution), the loss is approximately 84.3%. Early navigational charts for the Southport to Recherche Bay reveals extensive kelp beds on the offshore islands (Southport Island, Blanche Rock, George III Rock, Acteon Island, Blind Reef, the Images, South Break, Southeast Break) also extensive coastal beds in northern Southport, Recherche Bay, Rocky Bay and Fishers Point – Second Lookout Point. Most of these beds have been verified with recent field ground-truthing.



RM Comment: There is an interesting anomaly worthy of comment in Fig 3.7. The significant area of Giant Kelp beds described in the 1986 aerial survey appears to be ignored completely. The survey data clearly shows an area covering a phenomenal 12,975,123 m² in 1986 and 8,206,256 m² in 1999.

This is all the more interesting when considering the report's comment (Pg 97; Para 2)… Ground-truthing confirmed "beds at all the beds identified by recent Landsat TM and aerial survey". (Bold added to 'all' by RM)
.

• (Pg 97; Para 2): Recent aerial survey data (Aug_1999) recorded 16 beds in the region (totalling 8,206,256 m²), identifying a further 11 beds, including Stack of Bricks, Southport Bluff_north, Southport Island, Blanche Rock, George Third Rock, Eliza Point, Sullivans Point, Mary Ann Point (Recherche Bay), and in Rocky Bay (ie. Coalbins Bay, Waterhole Cove, Rocky Bay_south). The largest beds were recorded at Acteon Island (5,146,593 m²), Southport Bluff_north (1,071,833 m²), Images_Sullivan Point (434,502 m²), Southport Island (362,512 m²), George Third Rock (306,933 m²), and Blanche Rock (234,694 m²). However, there are very large discrepancies in some of the areal estimates of individual beds using aerial survey data (cf. with other mapping methods). By comparison, aerial photography consistently estimated less canopy area than aerial surveys, ie. Southport Bluff_north (288,682 m² in Dec_1988), Southport Island (131,681 m² in Feb_1990), Acteon Island (76,107 m² in Feb_1990), and George Third Rock (67,982 m² in Jan_1975). Significantly, ground-truthing (2001) has confirmed beds at all the beds identified by recent Landsat TM and aerial survey. While recent aerial photography (Feb_2001) identified only 2 beds (totalling 65,777 m²) at Pelican Island, the time of photography probably did not coincide with optimal kelp canopy formation.
.
RM Comment: The Southport to Recherche Bay example shows that there is an obvious selection of 'suitable' data to fit a specifically required outcome from this report. Even a cursory examination of the data used, tied in with the limitations and problems in the data as described, and taken in conjunction with comment within the text of the report show (at best) Giant Kelp bed natural fluctuations in size and growth.

Even ignoring the aerial survey data is irrelevant when considering the fact that the report says clearly that ground-truthing bore out the presence of all of the beds identified by the aerial survey.
.


---


.
Contradictory comment and anomalies in the report

There are contradictory themes that recur throughout the report. For example:

3.3.4: Aerial Photography vs Landsat TM Data

• (Pg 101; Para 5): In a few instances, the inclusion of the recent Landsat TM data resulted in different trends in kelp abundance. Over the past 50 years, Port Arthur showed an increase in kelp area (34%) with Landsat data, but a loss of kelp area (18%) with aerial photography. Similarly, Fortescue and North Bruny, both showed increases in kelp area (61% and 2%, respectively) over the past 20 years, based on aerial photography. However, inclusion of the Landsat data for these regions showed an overall loss of 55% and 13%, respectively.
RM Comment: Doesn't this demonstrate that the data is highly questionable, perhaps indicating that the aerial survey data may actually be more easily quantified?

3.3.5 Aerial Survey Data (1986, 1999)
.

• (Pg 101; Para 6): Overall, detailed temporal analysis of the aerial survey data is limited by the dataset, which includes only 2 sampling periods (June 1986 and August 1999). Further, the area north of Friendly Beaches was not surveyed in 1986. Despite these limitations, preliminary analysis indicates there has been a 23% total increase in the area of Macrocystis pyrifera (canopy area) along the east and south coast of Tasmania, between 1986-99, based on the aerial survey data. As such, the area of total kelp recorded in 1986 was 2,078 ha, compared to 2,562 ha in 1999.
.
RM Comment: This aerial survey data is an interesting addition to the other data sets. As is indicated above, ground-truthing verified the existence of all of the beds identified in the aerial survey for Southport to Recherche Bay.

Does this indicate in actual fact that the other data sets are of such limited or dubious value (due to the problems encountered) as to be highly inaccurate?

Reasons for this could be that with an aerial survey, the human eye (perhaps aided by ultra-violet sunglasses) can see more than a photograph that has problems with surface perturbation, reflectance, shadow, sun angle, cloud etc… To be pragmatic with this data, and if this has any degree of accuracy, the kelp beds have shown a "...23% total increase in the area of Macrocystis pyrifera (canopy area) along the east and south coast of Tasmania…" (See above).

Another recognised contradiction is described on page 102. This is very likely to be easily interpreted as natural fluctuations, as described previously.
.

• (Pg 102; Para 2): Of particular concern is the contradictory trends recorded for regions in close geographic proxity (sic). For instance, Freycinet recorded a 291% increase in kelp, while Schouten recorded a 54% decrease in kelp. Similarly, North Bruny recorded a 26% increase in kelp, while South Bruny recorded a 31% decrease in kelp
.
RM Comment: As alluded to previously, the introduced Japanese kelp Undaria pinnatifida is listed as a threat to Tasmanian Giant Kelps. This is perhaps another contradiction when 5.2.4 (Para 3) is taken into consideration:
.
5.2.4 (Pg 141; Para 5); Significantly, the results of manipulative, experimental field studies indicate that while Undaria pinnatifida is highly invasive, it does not displace or replace native algae, such as Macrocystis pyrifera. Rather, field results confirm that Undaria pinnatifida is disturbance-dependent, and colonises reefs after sea urchins have removed native algae. Importantly, this research also indicates that Macrocystis can establish and recover within Undaria forests, if there is no sediment or disturbance.
.
RM Comment: A further example of apparent contradictory statements within the report can be found in the description of evidence of increases in Giant Kelp beds between the Huon River and Southport (Page 94 on…)
.
• 3.3.2 (Pg 95; Para 4): Both, the recovery of beds (post-1970s) in the Lomas Point-Burnett Point region and the new beds in the lower Huon River, could be due to recent anthropogenic changes (ie. increased nutrients, stormwater runoff) in the lower Huon River and D’Entrecasteaux Channel. Kelp beds may have established and flourished in response to these anthropogenic changes, as kelp growth and reproduction is known to be strongly coupled with ambient nutrient availability and low temperature regimes
.
RM Comment: This seems to be at variance with the statement made in 1.4.1 where human (anthropogenic) discharges are listed as a threat to Macrocystis…
.
• 1.4.1: (Pg 28; Para 3) The distribution of Macrocystis is strongly influenced by natural oceanographic and physical influences, such as El Niño and La Niña episodes, storms, rainfall, and also, human-associated influences such as discharged wastewaters…



---

3.4.1 Major Episodes of Kelp Loss
.

• (Pg 103; Para 8): In examining the regional time-series of Macrocystis pyrifera distribution on the east and south coast of Tasmania (using historical aerial photography, Landsat TM and Alginates Australia data), several major episodes of kelp loss can be identified (see Table 3.23).

• (Pg 104; Para 1): In particular, major losses appear to have ocurred (sic) in the late 1960s, early 1970s, mid-1980s and early 1990s. These are discussed further in light of the major climatic, oceanographic and meteorological trends along the east of coast, in Section 4.4.3
RM Comment: So why was there such an outcry in the media by parties claiming that the report claimed or stated that the kelp "loss" was due to "overfishing"?

In fact the report never actually positively claims that overfishing is the cause of kelp 'decline' or 'loss' at all. It states that it is "possibly" or "probably" a contributing factor. These are very soft terms that allude to something, without making a definite statement that may need to be either qualified or quantified.

Chapter 4 deals with many major influences (physical processes) that affect Tasmanian Giant Kelp Forests, taking 28 pages in all. In contrast, 'Predator-Prey Interactions and Overfishing' is taken care of in 2½ pages (Pages 142 to 144). Comment on that part of the report follows directly.

The index to Chapter 4 is as follows (Pg 3):
.



5.2.5 (Pg 142) Predator-Prey Interactions and Overfishing
.
Increase in Sea Urchins
Populations of Centrostephanus rodgersii have increased off the east coast of Tasmania in the last 20 years, and were not recorded before 1978 (Andrew & Constable 1999). The reasons for this range extension are poorly understood but possibly relate to increased southerly penetration of the East Australian Current and associated warming of waters along the east coast (Edgar 1999). The native species of urchin, Heliocidaris has also increased significantly off the east coast over this time period (Will James, personal communication). (RM: It is worthy to bear in mind that this statement has no validation. There is neither verification nor validation from other sources that should be used here as well. This is simply an informed opinion.)

Importantly, over-exploitation of commercial abalone (Haliotis rubra) and lobster (Jasus edwardsii) can also contribute to the establishment of `urchin barrens’ (ie. areas denuded of seaweed). In New South Wales (where barrens occur over approximately 50% of nearshore reefs between Port Stephens and the Victorian border), reduction in abalone has possibly contributed to increases in the distribution and abundance of sea urchins, particularly Centrostephanus rodgersii (Andrew & Constable 1999). Further, C.rodgersii effectively prevents re-colonisation by abalone, with very few abalone found in `barrens’ habitats. Reducing densities of Centrostephanus can modify `barrens’ habitats however, to alternative habitats that can enhance the recruitment, survival and growth of abalone (Andrew 1993). Such a strong negative interaction between an established high-value fishery (abalone), and an under-exploited resource with considerable potential for development (sea urchins), offers a rare opportunity to develop complimentary (sic) fishery management plans.

Lobster (Jasus edwardsii) Fishery (Pg 142)
The southern rock lobster, Jasus edwardsii, is a major predator of sea urchins. Hence lobster abundance is likely to have a significant effect on urchin abundance and, indirectly, the abundance of Macrocystis pyrifera (via increased urchin grazing).

Observations of dense lobster populations in Marine Protected Areas and the catch rates of pioneer fishers suggest that lobsters were once extremely abundant on temperate reefs in Australia (Frusher et al. 1999). Adult lobsters are omnivorous and feed on a wide variety of organsims: crabs and other crustaceans, sea urchins, molluscs including bivalves, chitons and gastropods, and a variety of algal species. However, there is little doubt that lobsters were important predators of molluscs and sea urchins, the number of which in turn would have affected other aspects of the reef ecosystem, such as macroalgal cover. Further, recent lobster research in Tasmania has revealed that only adult (legal sized) lobsters are large enough to effectively prey upon urchins (Craig Johnson, University of Tasmania, pers.comm). Although, there has (sic) been no studies examining the direct effect of lobster abundance on prey species, the effect of reductions of lobster biomass on temperate ecosystems is likely to be significant. In many regions of Australia, the size of the harvestable portion of the population is thought to have been reduced by more than 90% (Frusher et al. 1999).

On the east coast, Jasus edwardsii stocks have been currently estimated at approximately 2- 3% of virgin biomass – thus, effectively removing lobster as a urchin key predator on inshore reef ecosystems. While further research is needed (particularly on the trophic and prey relationships of lobster), it is clear that the magnitude of the reduction in lobster biomass on the east coast, has probably been a significant contributory factor in the increased abundance of urchins in the region since the late 1970s. However, it is impossible to discern how much of the change in urchin abundance has been caused by intensive fishing, and how much is attributable to climate change.

.
RM Comment: The opening sentence of this last paragraph sounds suspiciously like a statement. This is somewhat inappropriate, because a statement as definite as this in a scientific paper should be backed up with a credible reference. There appears to be no referenced (published) basis for this far-fetched and unquantifiable statement. Who would be brave enough to declare the exact or even approximate number of a virgin biomass?

What brings an elevated degree of concern about this unfounded statement is a quote from Dr Edyvane published in The Bulletin (18 November 2003) in an article entitled "A cry for kelp", emotively authored by Taylor Bildstein and Bob Beale.

Dr Edyvane is quoted as saying "Lobster is an ecological keystone species on reefs and we are now down to less than 1% of the standing stock of the big lobsters…"
.
This was also presented in an article by the same authors, available at http://www.abc.net.au/science/news/stories/s987401.htm
.
Again there is no effort made to back the claim up with a reference, although in an article such as this, it is less likely to be needed. Further unquantified and disparaging comments about Tasmania's marine protection made in this (and another) article were soundly countered in a letter to the editor of 'The Mercury' by Professor Colin Buxton from the Tasmanian Aquaculture and Fisheries Institute (TAFI) in Hobart.

The sentence "However, it is impossible to discern how much of the change in urchin abundance has been caused by intensive fishing, and how much is attributable to climate change." is very clear on one thing. It indicates that to make the claim that overfishing is the cause of any supposed decline in Giant Kelp is to be extremely casual with the questionable scientific evidence and very economical with the truth. It is 'impossible' to know what the exact cause in increased urchin abundance is, and as such, it is yet to be determined what effect they have (either directly or indirectly) on Giant Kelp.

Blacklip Abalone (Haliotis rubra) Fishery (Pg 143)

Blacklip Abalone (Haliotis rubra) are some of the most abundant molluscs on rocky reefs in southern Australia. As a major grazer on rocky reefs in Tasmania (and southern Australia), the Blacklip Abalone, Haliotis rubra, plays a significant role in the ecology of temperate rocky reefs. As a competitor of sea urchins (particularly Centrostephanus rodgersii), it also has the potential to affect the distribution and abundance of large macroalgae, like Macrocystis pyrifera.

(Pg 144; Para 2): The trend in increased fishing effort for Haliotis rubra on the southeast and east coasts could possibly exacerbate increasing urchin abundance (particularly Centrostephanus rodgersii) in the region. This could have implications for the recovery and/or persistence of Macrocystis pyrifera habitat in this region.


RM Comment: There is no statement of fact or even a claim about overfishing here either. There are some pertinent comments made using soft terminology that obliquely indicate a potential or possible effect from single species fishing. Any sound claim that 'overfishing' is the cause of any perceived 'decline' in Macrocystis is entirely speculative if one is charitable, but more likely to be considered to be spurious and easily argued if there is a shred of realism within the discussion.

---

Threatened Species:

RM Comment: The purpose of the Edyvane report is to evaluate whether this species should be added to the list of threatened species. Considering the extreme variations in the methods used to gather data, and the stated doubts as to the validity or accuracy of that data as used in this report, the obvious conclusions to be drawn from the report are that:

• there is no clear scientifically validated evidence of Giant Kelp decline, rather it indicates substantial natural fluctuations as described anecdotally by commercial fishermen with decades of history in Tasmanian waters.
• there has been no evaluation of stunted Giant Kelp beds not visible from the surface.
• there is no effort in this project to make any subsurface evaluation of Giant Kelp beds.
• there are very questionable conclusions drawn from selected/ignored data.
• there are contradictory statements made within the text of the report.
• there are clear contradictory results from the various data collection methods.
• there are obvious questions raised as to species identification, especially from interpretation of black and white historic photographs that have been matched with more recent photographs of different resolution and scale. (Validation?)
• there is no real clear indication of specific causal links to Giant Kelp variations in population numbers and area, rather, generic assumptions are made from a number of unquantified anthropogenic effects or activities and/or natural oceanic processes.
• there is most certainly no consensus amongst steering committee members about Threatened Species listing for this species.
There is no clear or succinctly obvious reason for this species to be considered threatened. It has a massive natural variation and fluctuation in stock/population that is more easily born out from the nebulous data provided than the questionable conclusions from the report.

The Tasmanian Fishing Industry Council is vigorously opposed to this species being listed as a Threatened Species until convincing quantifiable scientific proof of any real decline or loss of Macrocystis pyrifera is produced.

If it could actually be clearly established that natural oceanic processes are causing a measurable decline in Macrocystis pyrifera, what protection would a Threatened Species listing achieve anyway?

---