Canada's fish habitat management is guided by the principle of "no net loss of the productive capacity of fish habitat" (NNL). Many development proposals are assessed using habitat information alone, rather than fish data. Because fish-habitat linkages are often obscured by uncertainty, uncertainty must be factored into NNL assessments. Using a quantitative framework for assessing NNL and lake habitats as a context, the implications of uncertainty for decision making are examined. The overall behaviour of a net change equation given uncertainty is explored using Monte Carlo simulation. Case studies from Great Lakes development projects are examined using interval analysis. The results indicate that uncertainty, even when large, can be incorporated into assessments. This has important implications for the habitat management based on NNL. First, schemas to specify relative levels of uncertainty using simple habitat classifications can support robust decision making. Second, attaining NNL requires greater emphasis on minimizing habitat loss and creating new areas to compensate for losses elsewhere and less on detailing small incremental changes in modified habitats where the fish response is difficult to demonstrate. Third, the moderate to high levels of uncer- tainty in fish-habitat linkages require that created compensation is at least twice the losses to reasonably ensure NNL. Resume : La gestion des habitats de poissons au Canada suit le principe selon lequel il ne doit y avoir « aucune perte nette (NNL) de capacite de production des habitats de poissons ». Plusieurs projets de developpement sont evalues d'apres les seules donnees de l'habitat, plutot que d'apres des donnees sur les poissons. Puisque les liens entre les poissons et les habitats sont souvent obscurs a cause de l'incertitude, cette incertitude doit etre prise en compte dans les evaluations NNL. L'utilisation d'un cadre quantitatif pour evaluer NNL dans le contexte d'habitats lacustres nous a permis d'examiner les implications de cette incertitude dans le processus de decision. Le comportement d'une equation de changement net en presence d'incertitude a ete etudie a l'aide d'une simulation de Monte Carlo. Des etudes de cas de projets de developpement dans le Grands-Lacs ont ete examinees a l'aide de l'analyse d'intervalles. Nos resultats indiquent que l'incertitude, meme importante, peut etre incorporee aux evaluations. Cela a d'importantes implications pour la gestion des habitats basee sur NNL. Premierement, des schemas servant a specifier le niveau d'incertitude a partir de classifications simples de l'habitat peuvent permettre des prises de decision sures. Deuxiemement, lorsque la reaction des poissons est difficile a demontrer, l'atteinte de NNL necessite une emphase plus grande sur la minimisa- tion des pertes d'habitat et sur la creation de nouveaux habitats pour compenser les pertes subies ailleurs, que sur de petites ameliorations successives dans les habitats modifies. Troisiemement, a cause des niveaux d'incertitude moyens a eleves qui existent dans les liens entre les poissons et les habitats, il faut que les habitats crees en compensation correspondent a au moins le double des pertes pour pouvoir raisonnablement assurer NNL. (Traduit par la Redaction) Minns and Moore 116
[1]
F. Yates,et al.
Sampling Methods for Censuses and Surveys
,
1950
.
[2]
G. A. Bradshaw,et al.
Conservation Ecology: Uncertainty as Information: Narrowing the Science-policy Gap
,
2008
.
[3]
Rashid Ahmad,et al.
A Distribution-free Interval Mathematical Analysis of Probability Density Functions
,
1975,
Interval Mathematics.
[4]
A. Keast,et al.
Nearshore fish habitat utilization and species associations in Lake Opinicon (Ontario, Canada)
,
1978,
Environmental Biology of Fishes.
[5]
R. G. Randall,et al.
Detecting the response of fish to habitat alterations in freshwater ecosystems
,
1996
.
[6]
C. K. Minns,et al.
Quantifying no net loss of productivity of fish habitats
,
1997
.
[7]
B. Cade,et al.
Estimating effects of limiting factors with regression quantiles
,
1999
.
[8]
P. Lawton,et al.
The Interface between Fisheries Research and Habitat Management
,
1996
.
[9]
Brian S. Cade,et al.
Modeling Stream Fish Habitat Limitations from Wedge-Shaped Patterns of Variation in Standing Stock
,
1996
.
[10]
C. K. Minns,et al.
Science for freshwater fish habitat management in Canada: Current status and future prospects
,
2001
.
[11]
K. Rose.
WHY ARE QUANTITATIVE RELATIONSHIPS BETWEEN ENVIRONMENTAL QUALITY AND FISH POPULATIONS SO ELUSIVE
,
2000
.
[12]
Francis Juanes,et al.
INFERRING ECOLOGICAL RELATIONSHIPS FROM THE EDGES OF SCATTER DIAGRAMS: COMPARISON OF REGRESSION TECHNIQUES
,
1998
.
[13]
Luther P. Aadland,et al.
Stream Habitat Types: Their Fish Assemblages and Relationship to Flow
,
1993
.
[14]
Yves T. Prairie,et al.
Evaluating the predictive power of regression models
,
1996
.
[15]
Max Henrion,et al.
Uncertainty: A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis
,
1990
.
[16]
J. Richardson,et al.
Distribution and abundance of freshwater fish in New Zealand rivers
,
1996
.
[17]
P. Fischer,et al.
Spatial distribution of littoral fish species in a large European lake, Lake Constance, Germany
,
1997
.
[18]
Ramon E. Moore.
Methods and applications of interval analysis
,
1979,
SIAM studies in applied mathematics.
[19]
Brean W. Duncan,et al.
SETTING RELIABILITY BOUNDS ON HABITAT SUITABILITY INDICES
,
2001
.
[20]
P. Auster.
Defining Thresholds for Precautionary Habitat Management Actions in a Fisheries Context
,
2001
.
[21]
J. Magnuson,et al.
Patterns in fish species composition across the interface between streams and lakes
,
2000
.
[22]
R. G. Randall,et al.
The relationship between an index of fish production and submerged macrophytes and other habitat features at three littoral areas in the Great Lakes
,
1996
.
[23]
Donald A. Jackson,et al.
Qualitative and quantitative sampling of lake fish communities
,
1997
.
[24]
R. G. Randall,et al.
A model simulating the impact of habitat supply limits on northern pike, Essox lucius, in Hamilton Harbour, Lake Ontario
,
1996
.