1. 1 Type-II Error, Precautionary Principle, and Accountability in Ecological Risk Assessment A new academic rule for pre- cautionary principle is needed. Statistical analysis is more important for environmental monitoring than it was. http://cod.ori.u-tokyo.ac.jp/~matsuda/2001/010830.ppt

2. 2 Precautionary principle Rio Declaration 1992, Principle 15 “In order to protect the environment, the precautionary approach shall be widely applied by States according to their capabilities. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost- effective measures to prevent environmental degradation.

3. 3 Convention on Biological Diversity JUNE 1992 “Noting also that where there is a threat of significant reduction or loss of biological diversity, lack of full scientific certainty should not be used as a reason for postponing measures to avoid or minimize such a threat,

4. 4 UN Framework Convention on Climate Change “Where there are threats of serious or ir- reversible damage, lack of full scientific certainty should not be used as a reason for postponing such measures, taking into account that policies and measures to deal with climate change should be cost-effective so as to ensure global benefits at the lowest possible cost.

5. 5 No academic rule for what we should say has been established. Scientists … Before the Earth Summit in 1992, –should give no comments to public without full scientific evidence; –keep their result irrespective of public opinion After must give some make their opinion a public consensus or win votes Galileo’s Inquisition

6. 6 IUCN Redlist Criteria (2001)

7. 7 Abuse of precautionary principle.. The extinction risk of a species is often uncertain. Threatened species is listed if one of the criteria A-E is satisfied. Therefore, it is to be listed solely because of rapid population decline, even though the extinction risk is low.

8. 8 G. Mace et al. 1992 : Species 19:16. (The validity of criterion A:) “ it can result in the listing of some species with very large, apparently secure populations ”. (Type-I error) “ However, linking [the rates of de- cline] to population size would exclude the listing of many populations with limited census data. ” (Type-II error)

9. 9 IUCN's credibility critically endangered. Mrosovsky N (1997) Nature 389:436 IUCN’s recommendations are based on sound and open science. Recent events suggest that this is not always the case.

10. 10 Allow criterion E to over-rule other criteria !? If we do not evaluate extinction risk, nobody disagreed with listing a species by criteria other than Criterion E. We disagreed with listing it by criteria A-D if estimated extinction risk is apparently low. No consensus was made in IUCN Marine Workshop. www.iucn.org/themes/ssc/redlists/marine/marine3.htm About 2/3 of IUCN Criteria Workshop participants disagreed with this option.

11. 11 How few evidence is necessary under precautionary principle? >5% of responsible scientists “stake” such pessimistic prediction. (I think no false prediction is charged with their honor) –#Scenarios is usually more than #scientists. >5% of type II error from an “unbiased” assumption. To describe what assumptions we used. ?? To need “consensus” or “majority” of untested assumptions.

12. 12 A Risk Management for sika deer H Matsuda, K Kaji, H Uno, H Hirakawa & T Saito (Official English web site by Hokkaido Government: http://www.hokkaido-ies.go.jp/HIESintro/ Natural/ ShizenHP2/SIKA/deerEng/

13. 13 History of deer in Hokkaido Island = Overexploitation, hunting bans and overabundance

14. 14 Damage on forests by deer http://www.marimo.or.jp/Kushiro_shichou/ezosika/cover.html

15. 15 Damage on agriculture and forestry in Hokkaido

16. 16 Adaptive management Adaptive management see also http://www.for.gov.bc.ca/hfp/amhome/AMDEFS.HTM We should –take into account of new data and knowledge (accountability). –change conservation actions with ecosystem state (adaptability). –make a falsifiable goal. EAJ guideline refers to monitoring only.

17. 17 Stage-Structured Model Stage-Structured Model (immature, ♂ & ♀ adults) N c, N f, N m : No. of calves(0.5), females & males( ≧ 1.5) L fc (t)= L mc (t)= exp[-Q(t)H c (t)]exp[-M c (t)]exp[-R c (t+1)], Estimate of Population in 1993 was 76- 164 thousands by aerial survey.

18. 18 Population parameters Age at maturity = 2 years !!! Sex ratio is 1:1 at birth, 1:3-1:5 at adults (forming harem) Female makes 1 calf every year! Estimate of Population in 1993 was 76-164 thousands by aerial survey. Rate of Natural Population Increase is 15%-20% per year!!! >interest of Salary loan or the merchant of Venice

19. 19 Density-dependent hunting pressure http://www.hokkaido-ies.go.jp/HIESintro/Natural/ShizenHP2/SIKA/DTdeerHP.htm %P>50%Emergency culling 25% < %P Gradual population reductions (catch females) 5% < %P Gradual population increases (catch males) %P <5% or after the severe winter Hunting bans

20. 20 2 Risk Management Matsuda et al. 1999 Res.Pop.Ecol. We set the upper and lower limit (%P + & %P - ) of population size P such that Prob{P<1000 individuals within the next 1 century} < 1% Prob{%P %P + } < 5%. Therefore, %P -  5%, %P +  50%

21. 21 1 Keep between 5% & 50% Relative Population Size

22. 22 Males disappeared in computer simulations!

23. 23 120,000 deer are under- estimation because… (Hokkaido Gov. 2000) Male adults were 20,000 -40,000 in 1993. (80,000 -160,000 )×20% ＝ (16,000 -32,000 ) >30,000 males have been caught every year since 1995. Despite it, male adults are still abundant. Good accountability

24. 24 3 Population Indices Catch & Watch per unit effort Spotlight census, Car & train accidents Damage of agriculture & forestry Uncertainty of absolute size

25. 25 The ratio estimator method (Cochran 1977), From ca.60 survey routes

26. 26 Population probably begin to decrease. 1993 95% CI Management Program started in 1998 Female hunting

27. 27 No parameters explain point estimate from spotlight census. r = # calves per mother r should be 0.4<r<0.5 Survival rate of ♀, ♂, calf

28. 28 Deviation between observed and simulated population size Chi-square distribution.

29. 29 Parameter values of “feasible runs” N(1993) should be 160,000 <N<320,000

30. 30 When do we stop Emergency culling of sika deer? Stakeholders (Farmers and Environmental NGOs) have different opinions and benefits. Statistical analysis of monitored population indices is definitely important.

31. 31 Revised Management Procedure of Whaling IWC/SC agreed with the RMP, but the IWC disagreed and continue to Moratorium of commercial whaling. The RMP monitors whale abundance, and commercial whaling is allowed if the abundance is >54% of initial size.

32. 32 Monitoring Abundance of Whales (Line Transect Method) >50% Population Decline with a small estimation error!

33. 33 Problem of Lead bullet fragments Shiretoko Museum http://www.ohotuku26.or.jp/shari/museum/home.htmlShiretoko Museum http://www.ohotuku26.or.jp/shari/museum/home.html

34. 34 The deer management threats endangered eagles The deer management threats endangered eagles Lead bullet fragments are consumed by scavenging eagles. 2,5,18 and 16 died Steller ’ s sea eagles were found from 1995 /96 to 1999 /00. Use of lead bullet in deer hunting was forbidden since 2000 /01 hunting season. But impact of lead on the extinction risk of eagles is uncertain.

35. 35 Individual-based model Individual-based model reanalysis of Ueta & Masterov (2000) Wild Bird Society of Japan Steller’s sea-eagles die until 20 years old. Hatchlings decreases by 75% with ¼ years 0000.11 0.7400000 00.7740000 000.931000 000 00 0000 N t+1 = NtNt

36. 36 Preliminary assumptions Preliminary assumptions by Ueta & Masterov (2000) Lead bullet contamination… –Increases mortality by 5% –Does not decrease reproductive success. Individual-based model investigates envi- ronmental & demographic stochasticity. Ignores non-lethal impact of lead Ignores increase of food-supply for eagles.

37. 37 Population decline of SS eagles Reanalysis of Ueta & Masterov (2000) Year after the present Average population size without contamination 5% mortality increase Cumulative extinction risk

38. 38 To detect extinction risk is difficult for long-lived species. Endocrine disrupters damage fertility, rather than mortality. In long-lived species, population trends are rarely apparent before the population falls into fatal. To detect cause of decline is difficult.

39. 39 TBTs-regulation － a successful story? Imposex: a female snail has male genitala and is sterile. Extremely high risk of extinction. All snails are to be affected by TBTs.

40. 40 Cause of imposex is TBTs in antifouling boat-bottom paints. Tributyltins contamination was associated with shell thickening oysters and imposex in sea snails TBTs were restricted in antifouling boat-bottom paints. Restrictive use of TBTs reduced TBT levels in environment and tissues of mussels and oysters.

41. 41 Does TBTs regulation pay? TBT-based paints save $6 billion for the world’s commercial fleet every year. TBT-based paints save $6 billion for the world’s commercial fleet every year.

42. 42 Ecosystem Services and Natural Capital Goods (Resource of Agriculture etc.)Goods (Resource of Agriculture etc.) Ecosystem services (Air, Water)Ecosystem services (Air, Water) Amenity, and intrinsic valuesAmenity, and intrinsic values is estimated to be in US$16-54 trillion per year, most of which is outside the marketis estimated to be in US$16-54 trillion per year, most of which is outside the market

43. 43 Conclusion 1.Precautionary scientific com- ments need some academic code based on statistical terms. 2.Statistical tools are indispensable in Decision-making among stakeholders under uncertain and precautionary actions.

44. 44 Kuril Population of Steller sea lions Population size Adults+juveniles (Perlov 1991) ca 80% decline sea lions and seals in Tyulen’i Island, 1992 Hokkaido Wakkanai Fisheries Experimental Station

45. 45 Lande & Orzack (1987 PNAS) Cumulative Probability Variance in decline rate

46. 46 Cumulative Extinction Probability for Steller’s sea lions (Takahashi & Matsuda in prep.) Mean time to extinction T=66 yr.

47. 47 Steller sea lions is listed in EN. Sea lions accumulate many sorts of environmental chemicals but biological cause of the decline is unclear. Magnitude of the decline is equivalent in effect to a 30-60% reduction in juvenile survival or 70%-100% reduction in female fecundity. The sea lions have been still caught by pest control in Japan.

48. 48 Limit of ecological risk assessment On-site assessment to the open ecosystem. 1 year assessment to the nonequilibrial ecosystem. ==“Hanshin Tigers Problem” No understanding of balance between succession and disturbance. ==“Sazae-san Syndrome”

49. 49 Factors of overabundance in sika deer. Conservation & Ban-on-hunting Loss of natural forests Increasing area of grassland No severe winter within the 25 years Extinction of wolves (ca1890)

50. 50 Population Dynamic Model incorporating uncertainty Rate of Natural population increase (λ) is between 15% and 20% per year. Survival rate L of calves: 46% ～ 54% L of female adults: 90% ～ 99% Reproductive success = 2λ(λ- L f )/ L c <10% yearly fluctuation Observation error of population <20%

51. 51 Should any few risk be avoided? (IWC/SC53 Annex D, 2001) Exploitation of whales with environmental variability was still “equivalent to an unsustainable ‘mining’” Unlike the RMP, “catch is assumed to be indefinitely constant and the time scales were far too long (10 45 years)” “the long time-scale was necessary to ex- amine the mechanisms of the interaction between environmental variability and exploitation.”