Difference between revisions of "Domestic fish consumption of the general population in Finland"
(technical edits and table 2 changed from kg/a to g/d like others) 
(headings updated and restructured) 

Line 1:  Line 1:  
[[Category:Exposures]]  [[Category:Exposures]]  
[[Category:Fish]]  [[Category:Fish]]  
−  {{variablemoderator=  +  {{variablemoderator=Jounistub=Yes}} 
==Question==  ==Question==  
−  What is the average fish consumption of individuals in the general Finnish population?  +  What is the average fish consumption of individuals in the general Finnish population? Imported fish may be considered. Sea and freshwater species may be considered separately. {{disclinkImported fish?}} 
==Answer==  ==Answer==  
−  +  == Rationale ==  
+  
+  === Data ===  
+  ==== 2007 study ====  
+  
+  Fish consumption by the population <ref name="Tero"> Tero Hirvonen. 2007. http://www.pyrkilo.fi/beneris/index.php/Image:Food_consumption_FINLAND.xls </ref> for twelve various fish species commonly consumed in Finland as normal distributions (weighed mean, weighed SD) and additionally indexed by age categories and gender.  
{ {{prettytable}}  { {{prettytable}}  
Line 41:  Line 46:  
}  }  
+  <anacode>5.717 5.897 11.55 12.96 29.61 39.02  
+  0.000e+000 0.000e+000 3.333 2.47 19.84 15.22  
+  7.858 7.6 8.884 10.2 37.08 49.41  
+  10.05 8.227 19.4 23.57 52.57 64.87  
+  3.636 2.402 3.684 3.106 26.13 32.6  
+  9.686 7.075 17.59 14.06 66.31 48.59  
+  0.663 1.339 4.328 4.142 19.94 26.41  
+  5.807 3.15 14.64 29.2 20.31 31.37  
+  10.09 7.458 16.48 14.44 54.51 32.06  
+  13.82 14.23 18.79 19.34 24.95 19.84  
+  18.65 18.83 22.52 17.77 0.000e+000  
+  7.981 6.591 7.733 8.539 47.43 39.26</anacode>  
+  
+  ==== Fisheries statistics 2006 ====  
−  +  There are two kinds of data. Actual consumption data and fishery catch data.<ref>Kalatalous tilastoina 2006 RKTL, page 23</ref> <ref name="RKTL">Kalatalous tilastoina 2006 RKTL, pages 3 and 9</ref> <ref>RKTL 2006(a) Estimates for filleting factors</ref> <ref>RKTL 2006(b) Estimates for factors of edible proportions</ref>  
+  
+  For the most important species (herring, salmon) the actual consumption estimates are available <ref>Methyl mercury: Kalatalous tilastoina 2006 RKTL, page 23</ref>, but for the most of the species consumption estimates are calculated from fishery catch. <ref>Methyl mercury: Kalatalous tilastoina 2006 RKTL, pages 3 and 9 </ref> For those species, fishery catch is first transformed to filleted fish weight <ref> Methyl mercury: RKTL 2006(a) Estimates for filleting factors</ref> and further with actual edible proportions <ref> Methyl mercury: RKTL 2006(b) Estimates for factors of edible proportions </ref>to fish consumption.  
+  
+  RKTL has data on overall fish consumption in Finland. This model yields a very similar estimate. {{disclinkLoss of fish during process?}}  
{ {{prettytable}}  { {{prettytable}}  
−  + '''Fish consumption as filleted weight (g/d)  +  + '''Fish consumption as filleted weight (g/d) (or kg/a?) 
! Fish species  Fish consumption  ! Fish species  Fish consumption  
    
Line 90:  Line 113:  
}  }  
−  ==  +  ==== Other data ==== 
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  
*[http://www.norden.org/pub/sk/showpub.asp?pubnr=2008:531&lang=1 TemaNord report on PCDD/F and PCB in Nordic countries]  *[http://www.norden.org/pub/sk/showpub.asp?pubnr=2008:531&lang=1 TemaNord report on PCDD/F and PCB in Nordic countries]  
−  
−  
The aim is to find probability distributions of daily consumption of various sea and freshwater fish species in six subgroups of the general Finnish population. The subgroups of interest are males and females in two age categories 02yr and 218yr, and adults (males and females combined) aged 1855yr and 55yr+. The data used to derive these distributions comes from <ref name="Tero" /> where the summary statistics (mean, sd and percentiles: 5th, 25th, 50th, 75th, 95th) of daily consumption of twelve fish species among fish users and information on the percentage of users are given. The data is provided separately for males and females in the following age groups: 1yr, 3yr, 69yr, 2534yr, 3544yr, 4554yr and 5564yr. There is a single data point of daily intake of whitefish by women aged 2534yr and 3544yr, and by males aged 4554yr. This information is not enough to create probability distribution. Thus, it was assumed that in each of these three sex/age groups the intake of whitefish is the same as intake of vendace. First, linear interpolation was used to approximate values of 5th, 25th, 50th, 75th, 95th percentiles of daily fish consumption among consumers in missing age/sex groups. For example, the ith percentile of daily consumption of species j in a group of 2 year old males was found by linear interpolation between points corresponding to ith percentiles of daily consumption of species j in groups of 1 year old and 3 year old males (for interpolation the mid age of every subgroup was used). The same technique was applied to age groups 45yr, 1013yr, 1417yr, 1824yr. The ith percentile of daily fish consumption by 0yr old children (users only) was approximated by the arithmetic mean of values of ith percentiles for months 7 through 11 obtained by linear interpolation between ith percentiles of fish consumption at age of 6 months and at age of 1yr (mid age of 1.5yr). Since young children don’t eat fish until they are 6 months old the fish consumption and the percentage of fish users at age of 6 months is zero. Moreover, it was assumed that the fish consumption and the percentage of fish consumers in age group 6574yr and in age group 75yr+ is the same as in group 5564yr. In the similar way the percentage of users of particular fish species in every missing age/sex group was estimated. Then, the available data (percentiles, mean and variance) was used to reconstruct the probability distributions of daily fish consumption among users in every age/sex group. The following procedure was used for that purpose. First, for a given fish species, age and gender the starting/initial cumulative distribution function (CDF) of daily consumption was constructed by linear interpolation between known percentiles. In general, this distribution will not reproduce the mean and sd as provided by the data. Therefore, the constrained optimization was further used to find the CDF that satisfies the percentiles given and whose mean and variance are as close as possible to the mean and variance from the data. For age groups for which the mean and variance were unknown only the first step of this procedure was applied. After that, information about the percentage of users was used to determine the probability distribution of daily consumption of particular fish species among all individuals in a certain sex/age group (users and nonusers). The resulting distributions were further grouped and combined (first by age and then, only in case of adults, by gender; population data can be found here <ref> Population of Finland years 1980  2007. http://en.opasnet.org/w/Image:Population_of_Finland_years_19802007.xls </ref>) to derive the fish consumption distributions in six target age/sex groups. Finally, since in <ref name="Tero" /> the fish habitat was not differentiated the species specific data on sea and inland intake proportions (assumed to be the same as proportion of sea and inland fishery catch and extracted from <ref name="RKTL" />) was used to estimate consumption of different sea and inland species by target subgroups.  The aim is to find probability distributions of daily consumption of various sea and freshwater fish species in six subgroups of the general Finnish population. The subgroups of interest are males and females in two age categories 02yr and 218yr, and adults (males and females combined) aged 1855yr and 55yr+. The data used to derive these distributions comes from <ref name="Tero" /> where the summary statistics (mean, sd and percentiles: 5th, 25th, 50th, 75th, 95th) of daily consumption of twelve fish species among fish users and information on the percentage of users are given. The data is provided separately for males and females in the following age groups: 1yr, 3yr, 69yr, 2534yr, 3544yr, 4554yr and 5564yr. There is a single data point of daily intake of whitefish by women aged 2534yr and 3544yr, and by males aged 4554yr. This information is not enough to create probability distribution. Thus, it was assumed that in each of these three sex/age groups the intake of whitefish is the same as intake of vendace. First, linear interpolation was used to approximate values of 5th, 25th, 50th, 75th, 95th percentiles of daily fish consumption among consumers in missing age/sex groups. For example, the ith percentile of daily consumption of species j in a group of 2 year old males was found by linear interpolation between points corresponding to ith percentiles of daily consumption of species j in groups of 1 year old and 3 year old males (for interpolation the mid age of every subgroup was used). The same technique was applied to age groups 45yr, 1013yr, 1417yr, 1824yr. The ith percentile of daily fish consumption by 0yr old children (users only) was approximated by the arithmetic mean of values of ith percentiles for months 7 through 11 obtained by linear interpolation between ith percentiles of fish consumption at age of 6 months and at age of 1yr (mid age of 1.5yr). Since young children don’t eat fish until they are 6 months old the fish consumption and the percentage of fish users at age of 6 months is zero. Moreover, it was assumed that the fish consumption and the percentage of fish consumers in age group 6574yr and in age group 75yr+ is the same as in group 5564yr. In the similar way the percentage of users of particular fish species in every missing age/sex group was estimated. Then, the available data (percentiles, mean and variance) was used to reconstruct the probability distributions of daily fish consumption among users in every age/sex group. The following procedure was used for that purpose. First, for a given fish species, age and gender the starting/initial cumulative distribution function (CDF) of daily consumption was constructed by linear interpolation between known percentiles. In general, this distribution will not reproduce the mean and sd as provided by the data. Therefore, the constrained optimization was further used to find the CDF that satisfies the percentiles given and whose mean and variance are as close as possible to the mean and variance from the data. For age groups for which the mean and variance were unknown only the first step of this procedure was applied. After that, information about the percentage of users was used to determine the probability distribution of daily consumption of particular fish species among all individuals in a certain sex/age group (users and nonusers). The resulting distributions were further grouped and combined (first by age and then, only in case of adults, by gender; population data can be found here <ref> Population of Finland years 1980  2007. http://en.opasnet.org/w/Image:Population_of_Finland_years_19802007.xls </ref>) to derive the fish consumption distributions in six target age/sex groups. Finally, since in <ref name="Tero" /> the fish habitat was not differentiated the species specific data on sea and inland intake proportions (assumed to be the same as proportion of sea and inland fishery catch and extracted from <ref name="RKTL" />) was used to estimate consumption of different sea and inland species by target subgroups.  
−  +  Unit: g/d  
−  +  === Calculations ===  
−  +  ==== Fisheries statistics 2006 ====  
−  
−  
−  
−  
−  
−  
−  
−  
−  
−  ===  
−  
−  
<anacode>Farmed_salmon_table+(Fish_onsumption_tabl/Population_of_fin)+Herring_consumption_ +(Wild_salmon_consumpt/population_of_fin)</anacode>  <anacode>Farmed_salmon_table+(Fish_onsumption_tabl/Population_of_fin)+Herring_consumption_ +(Wild_salmon_consumpt/population_of_fin)</anacode>  
Line 151:  Line 133:  
== References ==  == References ==  
−  <references />  +  <references /> 
−  
−  
−  
− 
Revision as of 11:47, 23 February 2021
Moderator:Jouni (see all) 
This page is a stub. You may improve it into a full page. 
Upload data

Contents
Question
What is the average fish consumption of individuals in the general Finnish population? Imported fish may be considered. Sea and freshwater species may be considered separately. D↷
Answer
Rationale
Data
2007 study
Fish consumption by the population ^{[1]} for twelve various fish species commonly consumed in Finland as normal distributions (weighed mean, weighed SD) and additionally indexed by age categories and gender.
Fish species  Consumption by 02 years old  Consumption by 218 years old  Consumption by 1855 years old 

Baltic herring (Clupea harengus membras)  5.717  11.55  29.61 
Herring (Clupea harengus)  0  3.333  19.84 
Vendace (Coregonus albula)  7.858  8.884  37.08 
Whitefish (Coregonus lavaretus)  10.05  19.4  52.57 
Pike (Esox lucius)  3.636  3.684  26.13 
Rainbow trout (Onchorhynchus mykiss)  9.686  17.59  66.31 
Shrimp (Pandalus sp.)  0.663  4.328  19.94 
Perch (Perca fluviatilis)  5.807  14.64  20.31 
Saithe (Pollachius virens)  10.09  16.48  54.51 
Atlantic salmon (Salmo salar)  13.82  18.79  24.95 
Pikeperch (Sander lucioperca)  18.65  22.52  0 
Tuna (Thunnus thynnus)  7.981  7.733  47.43 
<anacode>5.717 5.897 11.55 12.96 29.61 39.02 0.000e+000 0.000e+000 3.333 2.47 19.84 15.22 7.858 7.6 8.884 10.2 37.08 49.41 10.05 8.227 19.4 23.57 52.57 64.87 3.636 2.402 3.684 3.106 26.13 32.6 9.686 7.075 17.59 14.06 66.31 48.59 0.663 1.339 4.328 4.142 19.94 26.41 5.807 3.15 14.64 29.2 20.31 31.37 10.09 7.458 16.48 14.44 54.51 32.06 13.82 14.23 18.79 19.34 24.95 19.84 18.65 18.83 22.52 17.77 0.000e+000 7.981 6.591 7.733 8.539 47.43 39.26</anacode>
Fisheries statistics 2006
There are two kinds of data. Actual consumption data and fishery catch data.^{[2]} ^{[3]} ^{[4]} ^{[5]}
For the most important species (herring, salmon) the actual consumption estimates are available ^{[6]}, but for the most of the species consumption estimates are calculated from fishery catch. ^{[7]} For those species, fishery catch is first transformed to filleted fish weight ^{[8]} and further with actual edible proportions ^{[9]}to fish consumption.
RKTL has data on overall fish consumption in Finland. This model yields a very similar estimate. D↷
Fish species  Fish consumption 

Farmed salmon (sea+inland)  3.83 
Wild salmon  0.23 
Herring(sea)  1.92 
White fish(sea)  0.3 
Sprat(sea)  0.11 
Perch(sea)  0.44 
Flounder(sea)  0.01 
Pikeperch(sea)  0.2 
Bream(sea)  0.07 
Pike(sea)  0.35 
Vendace(sea)  0.06 
Burbot(sea)  0.02 
Wild salmon(inland)  0.45 
White fish(inland)  0.48 
Perch(inland)  1.25 
Pikeperch(inland)  0.27 
Bream(inland)  0.31 
Pike(inland)  1.53 
Vendace(inland)  1.38 
Burbot(inland)  0.16 
Other data
The aim is to find probability distributions of daily consumption of various sea and freshwater fish species in six subgroups of the general Finnish population. The subgroups of interest are males and females in two age categories 02yr and 218yr, and adults (males and females combined) aged 1855yr and 55yr+. The data used to derive these distributions comes from ^{[1]} where the summary statistics (mean, sd and percentiles: 5th, 25th, 50th, 75th, 95th) of daily consumption of twelve fish species among fish users and information on the percentage of users are given. The data is provided separately for males and females in the following age groups: 1yr, 3yr, 69yr, 2534yr, 3544yr, 4554yr and 5564yr. There is a single data point of daily intake of whitefish by women aged 2534yr and 3544yr, and by males aged 4554yr. This information is not enough to create probability distribution. Thus, it was assumed that in each of these three sex/age groups the intake of whitefish is the same as intake of vendace. First, linear interpolation was used to approximate values of 5th, 25th, 50th, 75th, 95th percentiles of daily fish consumption among consumers in missing age/sex groups. For example, the ith percentile of daily consumption of species j in a group of 2 year old males was found by linear interpolation between points corresponding to ith percentiles of daily consumption of species j in groups of 1 year old and 3 year old males (for interpolation the mid age of every subgroup was used). The same technique was applied to age groups 45yr, 1013yr, 1417yr, 1824yr. The ith percentile of daily fish consumption by 0yr old children (users only) was approximated by the arithmetic mean of values of ith percentiles for months 7 through 11 obtained by linear interpolation between ith percentiles of fish consumption at age of 6 months and at age of 1yr (mid age of 1.5yr). Since young children don’t eat fish until they are 6 months old the fish consumption and the percentage of fish users at age of 6 months is zero. Moreover, it was assumed that the fish consumption and the percentage of fish consumers in age group 6574yr and in age group 75yr+ is the same as in group 5564yr. In the similar way the percentage of users of particular fish species in every missing age/sex group was estimated. Then, the available data (percentiles, mean and variance) was used to reconstruct the probability distributions of daily fish consumption among users in every age/sex group. The following procedure was used for that purpose. First, for a given fish species, age and gender the starting/initial cumulative distribution function (CDF) of daily consumption was constructed by linear interpolation between known percentiles. In general, this distribution will not reproduce the mean and sd as provided by the data. Therefore, the constrained optimization was further used to find the CDF that satisfies the percentiles given and whose mean and variance are as close as possible to the mean and variance from the data. For age groups for which the mean and variance were unknown only the first step of this procedure was applied. After that, information about the percentage of users was used to determine the probability distribution of daily consumption of particular fish species among all individuals in a certain sex/age group (users and nonusers). The resulting distributions were further grouped and combined (first by age and then, only in case of adults, by gender; population data can be found here ^{[10]}) to derive the fish consumption distributions in six target age/sex groups. Finally, since in ^{[1]} the fish habitat was not differentiated the species specific data on sea and inland intake proportions (assumed to be the same as proportion of sea and inland fishery catch and extracted from ^{[3]}) was used to estimate consumption of different sea and inland species by target subgroups.
Unit: g/d
Calculations
Fisheries statistics 2006
<anacode>Farmed_salmon_table+(Fish_onsumption_tabl/Population_of_fin)+Herring_consumption_ +(Wild_salmon_consumpt/population_of_fin)</anacode>
See also
References
 ↑ ^{1.0} ^{1.1} ^{1.2} Tero Hirvonen. 2007. http://www.pyrkilo.fi/beneris/index.php/Image:Food_consumption_FINLAND.xls
 ↑ Kalatalous tilastoina 2006 RKTL, page 23
 ↑ ^{3.0} ^{3.1} Kalatalous tilastoina 2006 RKTL, pages 3 and 9
 ↑ RKTL 2006(a) Estimates for filleting factors
 ↑ RKTL 2006(b) Estimates for factors of edible proportions
 ↑ Methyl mercury: Kalatalous tilastoina 2006 RKTL, page 23
 ↑ Methyl mercury: Kalatalous tilastoina 2006 RKTL, pages 3 and 9
 ↑ Methyl mercury: RKTL 2006(a) Estimates for filleting factors
 ↑ Methyl mercury: RKTL 2006(b) Estimates for factors of edible proportions
 ↑ Population of Finland years 1980  2007. http://en.opasnet.org/w/Image:Population_of_Finland_years_19802007.xls