Versions

1.0 - 13-02-2019: Verson 1: First processing
1.1 - 15-12-2020: Major review of mapping from px-sources

Tables moved from G-drive to V-drive
Simplified from using S3 class constructors to and datatype checkups. Updates in DPLYR package (R-4.0) broke the constructor method.
Added data collection for non-metallic material used for concrete production. Used import values to calculate gravel/sand usage and validated against data from Steypustöðin.
Px-table titles for farming and fish statistics changed somewhat. Time consuming process.

1.2 - 27-09-2021: General review of datasources

Added a new datasource for graven pit usage and purchase of earthen material for road construction (mediocre solution)
Added a new datasource regarding fish farming production (px-web table). Need to see if this is a subtraction term for DE of fish.
Found an error in (SJA93031). Values were reported in kg rather than tonnes. Added a inline-validation for this error. The validation compares values from historic tables (not changing) and new data.

2.0 - 19-10-2021: Review of datasources

Added cross validation of export and DE of fish
Re-coding of helper functions. Removed “magic numbers” in functions and moved coefficients to DB. Added a BI tool to monitor the coefficient list.
Keep data on detail form within the process until the very end - move only aggregated data to DB
Added a fall-back mechanism where I fetch data from the V-drive. Now I can fetch data from the px-web if V-drive data is not found. V-drive data is more up-to date.
Added sanity checks for gravel, fish, and produce production.

2.1 - 08-07-2022: Update to harmonize with Food-flow calculations

Added DPO values from food processing into the calculations
Esimate RME coefficients for some fish products, IO-balance still off.

2.2 - 10-05-2023: Re-factoring of gravel and sand usage

Incorporated “sand and gravel v1” method from Eurostat.
Built a finance-to-quantity model for landscaping (bridges, roads, harbors)
Built a finance-to-quantity model for road construction. The model is mainly to validate data that I get (sometimes) from Vegagerðin.
Removed road-construction estimates from National Accounts

3.0 - 13-2-2024: Re-coding

Split the processing into independently executable sub-processes (Same method as in EPEA and EGSS)
Added data-quailty score into the processing.
Check the calulated data against last submitted data

What is the project

Table A in the EW-MFA module collects information on material flow from nature into the economy. This covers material independent if the collection is done by nationals or external economies. The material categories are:

MF1: Biomass: Weight of living animals and plants that are collected from land, lakes and ocean, not including farmed animals

MF11 All crops, assuming 15% humidity
MF12 All plants that are eaten by farm animals or fodder crops collected.
MF13 All wood collected.
MF14 All wild-caught fish and marine life. From manual “comprises all biomass extraction from non-cultivated (wild) natural biological resources”

MF2: Metals (obviously negligible here)
MF3: Non-metals earthen material, viz. sand, gravel, stones and other quarried material. Some ambiguity about if material is simply pushed into berms or flood-walls. Questionable if dams and shoreline barriers count unless they are transported from quarries
MF4: Plant material and oils for energy production

The data here is collected from published tables on farming and fishing. Specific data collection is needed to get information on wood (trivial amount) and non-metals (HUGE amount). Each data collection is stored in a sub-procedure and run in sequence when this script is executed. The data production includes estimates for data quality. In estimating the data quality the following score is given

If the value is used unaltered from the original datasource, the quality is 1
If the value is used unaltered from the original datasource, but split between two or more MFA categories, the score is lowered by factor of 0.75.
If the value is multiplied by a simple coefficient (e.g. by material density), the score is lowered by factor of 0.85. This is done even after a value is split between MFA categories.
If the value (e.g. monitary value) is used as an input into more complex model, (e.g. economic model), the score is the q-value of the model times 0.6
If the value is based on time-series modeling from previous years, the score is multiplied by 0.1

The quality score of an aggregate is calculated as the second moment average of the score, or the weighted average

The following processes are available/run

Keyrslur og skjöl fyrir töflu A
sub_proc	skjal_nafn	skjal_fannst
MF11	MFA-Table_A.mf11.rmd	TRUE
MF12	MFA-Table_A.mf12.rmd	TRUE
MF13	MFA-Table_A.mf13.rmd	TRUE
MF14	MFA-Table_A.mf14.rmd	TRUE
MF3	MFA-Table_A.mf30.rmd	TRUE
MF4	MFA-Table_A.mf4.rmd	FALSE
Table_A.vistun	MFA-Table_A.vistun.rmd	TRUE

MF.1 Biomass

MF11 Crops:

The principle data here are production values from the farming group. Published data is available on the PX-web under /Business sectors /Agriculture /Production. The data is also accessible on the Eurostat database and through FAO data repository, but the latter is generally less detailed. Validation of production is possible by exploring /Business sectors /Agriculture /Economic accounts of agriculture, but financial transactions of produce production is commonly on a different time-scale than the actual production.

Key datasources are therefore

LAN10103

Processing of harvest figures (LAN10103.px)

The table on the px-web with ID=LAN10103 contains key agriculture figures.

The data here is collected from internal datafile. Processing of the figures here returns figures to MF.1.1, mainly MF.1.1.1, MF.1.1.2, MF.1.1.7 and MF.1.2.1.1 and MF.1.2.1.1. The values for hay yield are given in cubic meters, rather than kilograms. Conversion from cubic meters to kg are found in the table ‘G:/Landbúnaður/AAP_verð og verðmæti/Framleiðslureikningar/EAA kerfi/2016/Lok-MP/03 - FORAGE PLANTS’.

Processing steps are as follows:

Fetch the data from the px-file: PX file found
Fetch the conversion keys between px titles and mfa keys from the db.table mf.gagnalindir_mfa_varpanir
Convert reported values from LAN10103 to MFA
Collect conversion factors from m3 to kg
Calculate kg

The processing is done in the following service functions

LAN10103.Table_A.check_keys: Collects keys from the px files and matches titles to db-keys.
LAN10103.Table_A.magntolur: Collects values from the px file and returns table with matched keys (umatched keys are ommitted). Ancillary argument here is “vinnsla_cd == ‘MASSI’ to indicate that only mass-figures are collected”
LAN10103.Table_A.rummal: Collects values from the px file and returns table with matched keys (unmatched keys are ommitted9. Ancillary argument here is “vinnsla_cd == ‘RUMMAL’. to indicated thast only volume-figures are collected. The function uses conversion factors to calculate mass from volume.
LAN10103.pxweb.Table_A.magntolur: Collects values directly from the px-online repository (used as a fallback mechanism if the px-file is not found on internal path).
LAN10103.Table_A.sanity_check. Collects the values and performs sanity check on the data (phased out 2020)

Px table stat:

Content summary of LAN10103.px
value.Ár	value.Tegund	value.value
2020 : 25	Kartöflur, tonn : 47	Min. : 27
2021 : 25	Ull, tonn : 47	1st Qu.: 380
2022 : 25	Egg, tonn : 47	Median : 1250
2014 : 22	Mjólk, tonn : 47	Mean : 311034
2015 : 22	Heyfengur alls, m3: 46	3rd Qu.: 37376
2016 : 22	Þurrhey, m3 : 46	Max. :3954181
(Other):676	(Other) :537	NA

Key check

Connected keys: 25
Unconnected keys: 1

Unconnected keys
ID	lind_cd	vinnsla_cd	lind_dalkur	eining_inn	hlutur	table_cd	mfa_cd	vara_cd	lind_texti	athugasemdir	vinnsluadferd_cd
NA	NA	NA	NA	NA	NA	NA	NA	NA	Korn alls, tonn	NA	NA

CHANGES 2024: Prior to 2023 the LAN10103 data only contained the item 'Korn, tonn' for all grain/wheat and other seed crops.
in 2023 the statice team decided to add the following items:

'Bygg, tonn'            - Barley
'Hafrar, tonn'          - Oats 
'Hveiti og rúgur, tonn' - wheat and rye
'Olíufræ, tonn'         - Oil seeds (mainly Canola)

and summarize 'Korn, tonn' values under 'Korn alls, tonn'

This changes the structure of the data slightly and introduces entry into MF.1.1.6 and reduces the crop in MF.1.1.1 somewhat

Data from the px-table:

Gögn úr LAN10103.magntolur()

The data here is a raw-data collection with minimal Dataquality issue. The DQ value is therefore = 1 for all values

Data to assembly

Staða á gagnasöfnum: (skref:1)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688	MF1221
MF11_landbunadur_magntolur	419	34	338791	MF112
MF11_landbunadur_magntolur	419	34	208	MF117
MF11_landbunadur_magntolur	419	34	148854	MF111
MF11_landbunadur_magntolur	419	34	3005	MF116

Combination of data and comparison to preivous submission

Last year’s submission (is_current) is collected for comparison with the data calculated here

## Warning: Removed 15 rows containing missing values (`geom_text()`).

message(‘MFA-Table_A.biti.mf11.rmd done’)

MF12: Crop residues (used)

Key datasources are:

LAN10103
LAN10102
LAN10201
LAN10205

NOTE:

In compilation 2018 all crops reported under "Korn, tonn" were allocated to crop-residue values (MF.1.2.2.1) . 
After meeting with field experts, this quantity was moved to MF.1.1.1 (edible) category
- Farmers are growing this crop predominantly for human consumption.
- In poor years most of the crop ends up as animal feed
- In some years, crop ends up in beer production or other similar products

It is therefore unjustified to allocate all the product to "scrap" category and
estimating success of the crop yield is an complication that reduces the data quality
in the processing.

MF121 Crop residues

MF1211 Straw

Here I will asssume thaht straw comes from harvesting grains and is the non-seed part of the harvest. Grain harvest is a very small quantity in Iceland, but Eurostat has guidelines for estimating the quantity of straw that comes from the harvest. The data here is therefore calculated directly from the quantity values in MF.1.1 above using products (vara) with the characteristics “vara_cd %in% c(‘BYGG’, ‘HVEITI’, ‘HAFRAR’)”. According this article, the portion of straw (= stems and leaves) represent half or more of the harvested oat plant. This is similar to what is used in the Eurostat compliation annexes. Therefore I will approximate the weight of the available straw as being the same as the amount of harvested grain for Oats, Flour and Barley.

Data to assembly

Staða á gagnasöfnum: (skref:2)
Vinnslueining	Fjöldi lína	Fjöldi tómra lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	0	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	0	32	219688	MF1211

MF1212 Other crop residues (sugar and fodder beats, beat leaves etc.)

Discussion with pig farmers in Iceland indicated that they receive trivial amounts of leaves from root crops (rutabagas, turnips and carrots), tubers (potatoes) and from brassica crop (cauliflower, broccoli and other similar). The values here will therefore be assumed to be zero for all years

MF122 Fodder crop and grazed biomass

Fodder crop (hey) should combine values that are collected to storage as well as the value consumed by animals on the field. Grass/hay was traditionally harvested and stored dry or in square bales in barns. Modernization in farming techniques now sees grass baled in round bales with added formic acid or other hydrolyzing agents.

The crop residue used here is reported in LAN10103 as volume data. The service function LAN10103.Table_A.rummal collects the values and performs the volume to mass conversion using acceptible conversion factors also used for FAO submissions. It can be assumed that dry-hay has lower density than wet-bales. The quantity of hay that is used is, however, likely to be significantly lower than quantity that is actually harvested (hay-bales are generally discarded and used as field-covering after 3 years)

MF1221 Fodder crops (including biomass harvest from grassland)

This is a sub-set of data from LAN10103 for “Heykögglar, tonn (m3)” and other items that indicate grass rather than food crops. The data is collected using the help-function “LAN10103.Table_A.rummal”

Data to assembly

Staða á gagnasöfnum: (skref:3)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221

Quality check of data against values in apro_cpnh1

Some of the data from LAN10103 are compatible to values submitted to Eurostat and published in the apro_cpnh1 datatable.

Structural changes in the Euorstat db in 2024 made this validation impossible/outdated. The original comments (icelandic) are below.

Tölur um uppskeru koma úr nokkrum px töflum sem eru unnar saman í flokka. Taflan “LAN10103” hefur þó hinsvegar nokkuð af tölum sem eru sambærilegar við Eurostat töfluna apro_cpnh1. Tölur hjá Eurostat eru með aðeins grófari einingu (Kílótonn) auk þess sem það vantar tölur fyrir 2011 í flesta flokka í Eurostat töfluna. Hins vegar þurfum við að gæta að því að tölurnar sem við gefum út hér séu sambærilegar, annars má gera ráð fyrir að við fáum einhverja athugasemd við skilin hjá okkur.

Uppvinnsla á Eurostat gögnum gerist í eftirfarandi skrefum:

Sæki töflurnar með því að nota Eurstat pakkann
Geymi lykla í gagnagrunninum í töflunni gagnalindir_mfa_varpanir. Söfnun og úrvinnsla á hvaða lyklar koma í Eurostat gögnunum eru unnar upp í pakkanum Eurostat_pakki með fallinu Export_Eurostat_keylist()
Keyri group-by til þess að setja Eurostat gögnin á sama form og eru í loka skilaskjalin
Safna töflum af gagnagrunninum og reikna upp samtöl eftir flokkum
Keyri outer-join á milli gagnasettana niður á ár og MF flokka.

Breytingar á eurostat grunninum gerði útaf við þessa vinnslu

MF1222 Grazed biomass

Here I need to estimate how much of the nature is eaten by the heard when it is not fed by products reported in MF.1.2.1. This happens for most of the summer months (e.g. from May to September) for Cows and Sheep, but Horses and Goats tend to be outside all year and fed over the wintermonths. Due to farming practices I also have to consider that the number of sheep on the land varies greatly throughout the year; Lambing season produces on average 1.6 lambs per sheep. The lambing season is in february-april and the heard is culled down in september-october. Here I will therefore build a bottom-up estimate for the plants consumed outside by

Estimate how many animals (per type) is fed outside (given that plants on the field are “edible” from May-September).
Estimate how many year-moths each animal type is eating outside
Use accepted coefficients for energy need per animal (by weight) to calculate energy need for the heard per months (note, Icelandic plants have nutrition content that is on average only 65% of the plant nutritients in warmer climates)

The number of animals that are eating outside over the summer months must be close to the number of animals registered on farms in November (registry data available) plus the number of animals that are slaughtered over and after the summer months (registry data avialable). The sense is that “winter slaugther” is minimal except for cattle and horses.

This means (as an example) that the number of sheep eating outside is:

\[ n\left(i=\textrm{SHEEP};\textrm{beit}\right) = n\left(i=\textrm{SHEEP};nóvember\right) + n\left(i=\textrm{SHEEP}; Slátrad \right) \]

A “veal” in slaughter figures must be less than 1 year old.
A “steer” is on average 2 years old
A “cow-other” may be 5-10 year old
A “gemlingur” is a summer-old sheep
A “gimbur” is less than 2 year old sheep/goat
A “sauður” is more than 2 year old sheep/goat, maximum 6 year old

This “behavior” is used to tweek the “n(i, slátrað)” value for each animal. Once the number of animals is established, I can estimate the mass of feed for the summer months according to:

\[ m(\textrm{feed}) = n(i;\textrm{beit})\times \frac{\textrm{mán.heim} - \textrm{mán.út}}{12}*\textrm{fóðurþörf}(i) \]

In order to estimate the values here I use four datatables:

LAN10102.px, which contains the number of animals alive in November (farm survey). The values here are inexact for horses as these animals are often owned by “gentlemen farmers”, households or semi-wild with unknown owner
LAN10205.px, which contains monthly animal slaugtered since 2012
LAN10201.px, which coontains the number of animals slaughtered per year since 1983 (used to fill data back to 1990 when LAN10205 is not available)
Fastar fyrir fóðurþörf dýra eftir tegundum (database build from FAO coefficients, MAST coefficients and from hours and hours (and hours) of discussions with specialists at MATÍS)

Summer here is considered June to October for sheep and horses. According to discussion with farmers and specialists sheep are usually fed hay until they are ferried to the highlands in June. Cows are fed year-long if they are producing milk, but get to “try and feed” outside over the summer (quote from Höskuldur Þráinsson, farmer on Miðhús, Vatnsdal)

Processing and mapping for number of animals alive in Nomvember (LAN10102.px)

This processing is:

Collect values from the table: px found
Check if the same table exists on the web-db: Búpeningur eftir landsvæðum frá 1980: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The processing uses the following service functions

LAN10102.Table_A.check_keys: Collects the keys and checks what keys are connected and which ones are not connected
LAN10102.Table_A.bupeningur_i_november: Returns data from the datatables which have matched keys.
LAN10102.Table_A.bupeningur_i_november.web: Fallback function if the internal px table is not found

the px table summary

Summary of LAN10102.px
value.Landsvæði	value.Búpeningur	value.Ár	value.value
Allt landið :595	Nautgripir alls: 218	1998 : 106	Min. : 1.0
Suðurland :304	Mjólkurkýr : 218	1999 : 103	1st Qu.: 766.2
Norðurland eystra :300	Kvígur : 218	2000 : 102	Median : 3857.0
Höfuðborgarsvæði og Suðurnes:299	Geldneyti : 218	2001 : 101	Mean : 34752.3
Austurland :297	Sauðfé alls : 218	2002 : 101	3rd Qu.: 24705.0
Norðurland vestra :295	Ær : 218	2004 : 101	Max. :827927.0
(Other) :532	(Other) :1314	(Other):2008	NA

Key check

Land district-connected keys: 8
Land district-not-connected keys: 0

Here I use “hlutur=1” as a filter to pick up the data and map it to the processing. “hlutur=0” is a “don’t use” filter. Here I don’t need/intend to do analysis on NUTS regions, so “Allt landið” is the appropriate pick

Animals-mapped: 15
Animals-not mapped: 0

Here I use “hlutur” as a filter for animals that are actually eating outside during the summer months. This ommits minks, fox, pigs and other non-grazing animals

Data from the px file

Gögn úr LAN10102.Table_A.bupeningur_i_november()

Processing of th enumber of slaughtered animals (LAN10205.px)

This processing is identical to above:

Collect values from the table: px table found
Check if the same table exists on the web-db: Kjötframleiðsla eftir mánuðum frá 2012: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The processing uses the following service functions

The processing is stored in the following service functions:

LAN10205.Table_A.check_keys: Collects titles in the px table and check mapping
LAN10205.Table_A.bupeningur_slatrad: Collects values from the px table and returns figures that are mapped (ommits valuse without mapping)
LAN10205.pxweb.Table_A.bupeningur_slatrad: Fallback to web-table if the internal px table is not found

px table summary

innihald LAN10105.px
value.Eining	value.Tegund	value.Mánuður	value.value
Kíló :882	Dilkar :154	2012M01: 14	Min. : 1
Fjöldi:882	Fullorðið sauðfé:156	2012M03: 14	1st Qu.: 960
NA	Svín :292	2012M04: 14	Median : 8588
NA	Alifuglar :292	2012M05: 14	Mean : 245407
NA	Kálfar :290	2012M08: 14	3rd Qu.: 413223
NA	Ungnaut :290	2012M09: 14	Max. :4892360
NA	Kýr :290	(Other):1680	NA

Mapping check

Units-mapped: 2
Units-not mapped: 0

Here I use “hlutur=1” to indicate which of the measure I am using for the processing

Animals-mapped: 7
Animals-not mapped: 0

Here I use “hlutur=1” to indicate which animals are used for processing. There is an option to use “hlutur<1” to indicate if the animals are older than 1 year old when they come in for slaughter, but this option is not as flexible as using a mapping table

Processing and data outcome

The processing here is slightly more involved since the time variable here is year+months. I therefore leave the option to use the variable f.agg=TRUE to calculate the average over the year, whereas f.agg=FALSE returns the monthly values. Processing-wise it would be more convenient here to put the “end-of-year” for the producing farms in November, rather than in December, since we have the november “end of year” figures. Analysis show, however, that the number of animals slaugthered from November-December is trivial compared to number of animals slaughtered from June-November.

The data shows that the most dangerous months for being a sheep is September-November. The mortality rate of other animals is also lower in these months (probably since the slaughterhouses are busy). Figures here are not exact for poultry and fur-producing animals, since farmers are apparently allowed to establish their own slaughterhouses.

Processing of pre 2012 figures (LAN10201.px)

This processing is identical to above:

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Kjötframleiðsla eftir tegundum frá 1983: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The processing is done by the following service functions:

LAN10201.Table_A.check_keys: Collects titles from the table and checks for mapping
LAN10201.Table_A.bupeningur_slatrad: Collects values from the px-file and returns table with mapped values.
LAN10201.pxweb.Table_A.bupeningur_slatrad: Fallback function for processing the data from the px-web if the internal document is not found.

px table summary

summary of LAN10201.px
value.Landsvæði	value.Búpeningur	value.Ár	value.value
Allt landið :595	Nautgripir alls: 218	1998 : 106	Min. : 1.0
Suðurland :304	Mjólkurkýr : 218	1999 : 103	1st Qu.: 766.2
Norðurland eystra :300	Kvígur : 218	2000 : 102	Median : 3857.0
Höfuðborgarsvæði og Suðurnes:299	Geldneyti : 218	2001 : 101	Mean : 34752.3
Austurland :297	Sauðfé alls : 218	2002 : 101	3rd Qu.: 24705.0
Norðurland vestra :295	Ær : 218	2004 : 101	Max. :827927.0
(Other) :532	(Other) :1314	(Other):2008	NA

Mapping check

Units - mapped: 5
Units - not mapped: 1

Titles not mapped í mfa.gagnalindir_mfa_varpanir
ID	lind_cd	vinnsla_cd	lind_dalkur	eining_inn	hlutur	table_cd	mfa_cd	vara_cd	lind_texti	athugasemdir	vinnsluadferd_cd
NA	NA	NA	NA	NA	NA	NA	NA	NA	Seld innanlandsframleiðsla í tonnum	NA	NA

Animals - mapped: 5
Animals - not mapped: 0

Processing of values from px table

This processing is less involved than the processing of the monthly values from 2012.

Gögn úr LAN10201.Table_A.bupeningur_slatrad()

In the final processing I discard values post 2012

Combination and final processing:

The processing uses one service function:

Table_A.lifmassi_af_beitarlondum, which uses
- LAN10102.Table_A.bupeningur_i_november,
- LAN10201.Table_A.bupeningur_slatrad og
- LAN10205.Table_A.bupeningur_slatrad

The calculation returns the number of animals over thesummer, the length of the summer and the biomass that is consumed.

source(file.path(paths$source_table_a, 'Table_A.lifmassi_af_beitarlondum.R'))

tmp_path.list = list(
    bupeningur_i_husi.path = paths$LAN10102$path,
    bupeningur_slatrad_eldri.path = paths$LAN10201$path,
    bupeningur_slatrad_yngri.path = paths$LAN10205$path
  )

tmp_lifmassi_af_beitarlondum <- Table_A.lifmassi_af_beitarlondum(
  paths = tmp_path.list,
  this.connection = paths$db())|>
  filter(ar >= 1990)

The processing in Table_A.lifmassi_af_beitarlondum() does the following:

Collect data from the px tables
Combines valuse from the older collection and the new table.
Calculates the number of slaugtered animals (group-by process)
Collects number of animals in November by the same groups.
Calculates the number of animals per months from June to November and estimates weight of lambs per month
Collects the food requirements from mfa.likan_grazed_biomass_coefficients
Uses the required feed per year for the animal
Scales the required food down to month for the animals over summer
Connects the mfa-mappings from mfa.gagnalindir_mfa_varpanir using the switch lind_cd = ‘LIFMASSI_BEITARLOND’
Returns the data

The data here is given a lower data-quality measure depending on number of estimates and steps in calculation. Results:

Gögn úr LAN10201.Table_A.bupeningur_slatrad()

Data for the assembly

Staða á gagnasöfnum: (skref:4)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221
MF1222_lifmassi_af_beitarlondum	66	33	8504456	MF1222

Combination of data and comparison to preivous submission

Last year’s submission (is_current) is collected for comparison with the data calculated here

## Warning: Removed 105 rows containing missing values (`geom_text()`).

COMPILATION 2024: data for hay was not available in February

## MFA-Table_A.mf12.rmd done

MF13 Wood

While Icelanders are quite proud of the shrubbery that they consider to be woods, this greenery is properly classified as “shrubbery” or “swampwood” according to Eurostat.

Some forestry services in Iceland fund supplement their operation by selling firewood, or even occational planks. There is also some christmastree farms that allow the general population to cut their own trees as a part of the family holiday traditions. The total sales from bake-sales is, however, close to being equal to the actual sales of christmas trees. This factor is there not expected to be a substantial part of any material flow statistics, but available data is collected herein.

MF13.1 Timber (Industrial roundwood)

FAO defines industrial roundwood as logs that are sawn into boards or peeled into plyes for plywood. These trees should measure between 2 and 6 inches in diameter, which instantly limits the number of available trees for this extraction in iceland

MF.1.3.2 Wood fuel and other extraction

This contains

Christmas trees
Branches used as industrial carbon source
Small logs sold as firewood

processing

Here I pick up the data from the FAO database, rather than trying to hunt down the data internally.

Default gildi fyrir timburframleiðslu, vonandi koma betri tölur seinna
table_cd	mfa_cd	ID_Gagnalind_mfa_vorpun	ar	gildi	eining_inn	vara_cd	lind_texti	reiknad_dags
Table_A	MF132	1	2011	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2012	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2013	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2014	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2015	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2016	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2017	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2018	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2019	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2020	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2021	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2022	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2023	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09
Table_A	MF132	1	2024	3	TONN	TIMBUR	Default gildi fyrir viðarframleiðslu á Íslandi (vonandi tímabundið)	2024-04-09

Information from FAO

The values are provided by Jón Guðmundur Jónsson, which submits the data to FAO. The data is collected from the FAO database site and stored locally in *C:/Users/ThorsteinnA/OneDrive - Public Administration/Documents/R/EW-MFA/_GognInn/gogn_timbur*. The datafile is processed. Each file has a timestmp built into it’s name. The newest file is processed.

file	date	newest
FAOSTAT_data_10-18-2021.csv	2021-10-18	FALSE
FAOSTAT_data_2-20-2020.csv	2020-02-20	FALSE
FAOSTAT_data_en_2-16-2024.csv	2024-02-16	TRUE

The data (production only) is as follows (in m3)

Here the recovered paper is included (obviously not a product of the forestry industry). The following mapping is used to process the data.

Mapping fro product descripion to MFA categories
Item	density	use	mfa_cd
Wood fuel, coniferous	820	TRUE	MF132
Wood fuel, non-coniferous	900	TRUE	MF132
Sawlogs and veneer logs, coniferous	820	TRUE	MF131
Sawlogs and veneer logs, non-coniferous	900	TRUE	MF131
Other industrial roundwood, coniferous (production)	820	TRUE	MF131
Other industrial roundwood, non-coniferous (production)	900	TRUE	MF131
Wood charcoal	3500	TRUE	MF132
Wood chips and particles	800	TRUE	MF132
Wood pellets	800	TRUE	MF132

Staða á gagnasöfnum: (skref:5)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221
MF1222_lifmassi_af_beitarlondum	66	33	8504456	MF1222
MF13_FAO_timbur	24	12	22237.94, 10630.10	MF131, MF132

Extending the FAO data

The time on the FAO data is a trouble here, as well as the number of datapoints that is imputed or estimated. The data is, however, not available on the PX site at statistics Iceland. The newest year in the FAO data is 2022, and in some cases we want to extend the data to the previous or current year. Here I chose to use lvcf to impute the data.

##                                 Length Class  Mode
## MF11_landbunadur_magntolur      10     tbl_df list
## MF1211_landbunadur_stra         10     tbl_df list
## MF1221_landbunadur_rummal       10     tbl_df list
## MF1222_lifmassi_af_beitarlondum 10     tbl_df list
## MF13_FAO_timbur                 10     tbl_df list
## MF13_FAO_ext                    10     tbl_df list

Combination of data and comparison to preivous submission

Last year’s submission (is_current) is collected for comparison with the data calculated here

## Warning: Removed 4 rows containing missing values (`geom_text()`).

## $MF13_FAO_timbur
## # A tibble: 24 × 10
##    table_cd mfa_cd ID_gagnalind_mfa_vorpun    ar gildi eining_inn vara_cd
##    <chr>    <chr>                    <dbl> <int> <dbl> <chr>      <chr>  
##  1 Table_A  MF131                        1  2011 2034. TONN       TIMBUR 
##  2 Table_A  MF132                        1  2011  834  TONN       TIMBUR 
##  3 Table_A  MF131                        1  2012 2241. TONN       TIMBUR 
##  4 Table_A  MF132                        1  2012  754. TONN       TIMBUR 
##  5 Table_A  MF131                        1  2013 2248. TONN       TIMBUR 
##  6 Table_A  MF132                        1  2013  760. TONN       TIMBUR 
##  7 Table_A  MF131                        1  2014 2248. TONN       TIMBUR 
##  8 Table_A  MF132                        1  2014  760. TONN       TIMBUR 
##  9 Table_A  MF131                        1  2015 2248. TONN       TIMBUR 
## 10 Table_A  MF132                        1  2015  760. TONN       TIMBUR 
## # ℹ 14 more rows
## # ℹ 3 more variables: lind_texti <chr>, vinnsla <chr>, reiknad_dags <date>
## 
## $MF13_FAO_ext
## # A tibble: 4 × 10
##   table_cd mfa_cd ID_gagnalind_mfa_vorpun    ar gildi eining_inn vara_cd
##   <chr>    <chr>                    <dbl> <int> <dbl> <chr>      <chr>  
## 1 Table_A  MF131                        1  2023  2722 TONN       TIMBUR 
## 2 Table_A  MF131                        1  2024  2722 TONN       TIMBUR 
## 3 Table_A  MF132                        1  2023   836 TONN       TIMBUR 
## 4 Table_A  MF132                        1  2024   836 TONN       TIMBUR 
## # ℹ 3 more variables: lind_texti <chr>, vinnsla <chr>, reiknad_dags <date>

## MFA-Table_A.mf13.rmd done

MF14: Wild fish catch, aquatic plants and animals, hunting and gathering

Given that a third of the Icelandic economy is based around fishing, this is obviously a relative part in the data. This processing is connected with the processing developed for calculating the actual edible fish and generation of products to export or consumption within iceland.

MF141 Wild fish

Statistics Iceland has numerous publications for fising. I also have access to the internal fish-statistics database that allows me an unparallelled access to weighting data of fish landings. The database data is, however, significantly cleaned before publication, so the process here for MFA is still based on px-data.

The main datasources are:

SJA09005: Catch of Icelandic vessels from all fishing areas
SJA09031: Catch and values by species, type of landing and months
LAN10302: Salmon caught in rivers and lakes
LAN10301: Salmon caught in wild-rivers (used to calculate the average weight )
LAN10303: Hunting and gathering

Processing of SJA09005, Catch on sea

SJA09005 is a simplified collection of the key catch figures. This includes reporting of total catch of the icelandic fishing fleet. It is not completely clear that this does not include the catch of foreign fishing fleet that does not land fish in Icelndic ports.

Processing here is similar to other px tables

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Afli íslenskra fiskiskipa af öllum miðum eftir fisktegundum 1945-2022: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

A specific worry here is a double counting error that I found in 2021 processing.

The processing uses the following service functions

SJA09005.Table_A.check_keys: Checking keys
SJA09005.Table_A.afli_ur_sjo: Collect data

px table from local drive

content SJA09005.px
value.Fisktegund	value.Ár	value.value
Heildarafli: 78	1962 : 9	Min. : 1077
Þorskur : 78	1963 : 9	1st Qu.: 40457
Ýsa : 78	1964 : 9	Median : 84700
Ufsi : 78	1965 : 9	Mean : 267731
Karfi : 78	1966 : 9	3rd Qu.: 294030
Síld : 78	1967 : 9	Max. :2199111
(Other) :200	(Other):614	NA

The data reaches back to 1945, which is way outside of what is needed. Ther eis alos a category called “Heildarafli” which does include other fish types that are not in the breakdown.

Check mapping

Fish type - mapped: 9
Fish type - not mapped: 0

Data from the table

##    vara_cd            table_cd            mfa_cd           lind_texti       
##  Length:33          Length:33          Length:33          Length:33         
##  Class :character   Class :character   Class :character   Class :character  
##  Mode  :character   Mode  :character   Mode  :character   Mode  :character  
##                                                                             
##                                                                             
##                                                                             
##   eining_inn              ar           gildi           vinnsla         
##  Length:33          Min.   :1990   Min.   :1021020   Length:33         
##  Class :character   1st Qu.:1998   1st Qu.:1153683   Class :character  
##  Mode  :character   Median :2006   Median :1414598   Mode  :character  
##                     Mean   :2006   Mean   :1478979                     
##                     3rd Qu.:2014   3rd Qu.:1712252                     
##                     Max.   :2022   Max.   :2199111                     
##   reiknad_dags       
##  Min.   :2024-04-09  
##  1st Qu.:2024-04-09  
##  Median :2024-04-09  
##  Mean   :2024-04-09  
##  3rd Qu.:2024-04-09  
##  Max.   :2024-04-09

Gögn úr SJA09005.Table_A.afli_ur_sjo()

Sanity check of the values

I have had several issues with the published records here, so I built a quick sanity check. Reasonable values are:

The catch is unlikely to be less than million tonnes per year
The catch is unlikely to be avbove three million years
I should only have a single value per year

name	items	passes	error	warning	expression
Lower.limits	33	33	FALSE	FALSE	gildi - 1e+06 >= -1e-08
Upper.limit	33	33	FALSE	FALSE	gildi - 3e+06 <= 1e-08
Unique	33	33	FALSE	FALSE	is_unique(mfa_cd, ar)

Values for the collection

This table has very little detail that allows us to separate the catch down to the different MFA categories. We defninitely have more detailed data.

Processing of: SJA09031; Catch and values by species, type of landing and months:

This table has slightly different breakdown from SJA09005, but here we get better breakdown by type of catch, including crustaceans and other types of animals that should be classified into MF142 in stead of MF142

Processing here is similar to other px tables

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Afli og verðmæti afla eftir tegund löndunar og fisktegundum 1982-2022: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The table has more dimensions that need to be mapped for the processing here. This means that we map

The catch “Hákarl”, “Háfur”, “Humar”, “Rækja” are mapped to MF142: All other aquatic animals and plants
Other types that are not “totals” are mapped to MF141
The landing designation “Landað erlendis til bræðslu” og “Selt úr skipi erlendis” er omitted, since this indicates that the fish is not broght to shore in Iceland. This catch should be reported in the trade statistics.
The unit “Magn” is chosen
The month “Alls” is chosen.

px tafla

innihald SJA09031.px
value.Mánuður	value.Ár	value.Eining	value.Löndunartegund	value.Fisktegund	value.value
Alls :15919	2002 : 4675	Aflamagn (tonn) :77657	Allar löndunartegundir:34797	Allar fisktegundir: 7321	Min. : -33
október :11941	2003 : 4656	Verðmæti (1000 kr):77394	Til vinnslu innanlands:31444	Þorskur : 6602	1st Qu.: 19
september:11885	2004 : 4561	NA	Á fiskmarkaði :23881	Langa : 6094	Median : 402
maí :11860	1992 : 4555	NA	Sjófryst :22714	Ýsa : 6058	Mean : 176352
mars :11779	2000 : 4543	NA	Í gáma til útflutnings:21074	Ufsi : 6026	3rd Qu.: 8522
ágúst :11655	1998 : 4520	NA	Selt úr skipi erlendis: 7695	Karfi : 5986	Max. :195005151
(Other) :80012	(Other):127541	NA	(Other) :13446	(Other) :116964	NA

Check mapping

Catch: Mapped: 41
Catch: Not mapped: 1

Catch titles that are not mapped (add to mfa.gagnalindir_mfa_varpanir)
ID	lind_cd	vinnsla_cd	lind_dalkur	eining_inn	hlutur	table_cd	mfa_cd	vara_cd	lind_texti	athugasemdir	vinnsluadferd_cd
NA	NA	NA	NA	NA	NA	NA	NA	NA	Allar fisktegundir	NA	NA

Here I use “hlutur=1” to indicate that the catch should be used

Landing area: Mapped: 10
Landing area: Not mapped: 1

Landing area titles that are not mapped (add to mfa.gagnalindir_mfa_varpanir)
ID	lind_cd	vinnsla_cd	lind_dalkur	eining_inn	hlutur	table_cd	mfa_cd	vara_cd	lind_texti	athugasemdir	vinnsluadferd_cd
NA	NA	NA	NA	NA	NA	NA	NA	NA	Á fiskmarkaði	NA	NA

Here I use the mfa_cd key to indicate if the value should be used rather than using “hlutur”

Unit of measure: mapped: 2
Unit of measure: not mapped: 0

Compilation 2024: The units of measure had changed (yet again). This needs to be monitored. 
Titles for the catch weight have changed from "Magn"->"Tonn"->"1000 kg"-> "Aflamagn (tonn)"
Titles for the catch value have changed from "Verðmæti"->"1000 kr"->"Verðmæti (1000 kr)"

Here I use the mfa_cd to indicate which of the UOM should be used,

Compilation 2024: Due to multiple versions of the table I had mapped the same value twice for "lind_texti" to a key. This would have meant that all values would be doubled

Months: Mapped: 13
Months: Not mapped: 0

Here I chose to take only the “TOTAL” for calculation since I am not as concerned with monthly variations. The monthly variations are used in an alternative processing (AFLI_LANDAD_VINNSLA) that I do elsewhere for cross validation between export and catch amount

The values

The service funtion for this processing is SJA09031.Table_A.afli_landad(). This function does the following

Gets the px table from the local drive (fallback web)
Collects mapping using the SJA09031.Table_A.check_keys() function
Runs a key-join and omitts values that have either hlutur=0, or has mfa_cd = NULL.
Returns the data.

In earlier processing I had the function also run a group-by on the data and return the sum, but in later development I found it better to keep the data with full detail and run the summary later.

Gögn úr SJA09031.Table_A.afli_landad()

This data has similar structure/shape as SJA09005, but here I also have the MF142 data.

Validate check on the data

Since I have the fish-type in here, I could apply more validation rules, like:

Catch is rarely over 300 thousand tonnes per years for a single species, except for Capelin

It would be useful to see if the swings in catch is normal in the data.

name	items	passes	fails	nNA	error	warning	expression
Below.300.kt	1280	1245	35	0	FALSE	FALSE	gildi - 3e+05 <= 1e-08 & vara_cd != “LODNA”
Unique.values	1280	964	316	0	FALSE	FALSE	is_unique(mfa_cd, ar, vara_cd, lind_texti)

Here I need to re-check the unique constraints on the data

Validate rules for fish (MF141)

Catch in MF141 is more than million tonnes, but less than three million tonnes

name	items	passes	fails	nNA	error	warning	expression
Yfir.miljón	33	27	6	0	FALSE	FALSE	gildi - 1e+06 >= -1e-08
Undir.þremur.miljónum	33	33	0	0	FALSE	FALSE	gildi - 3e+06 <= 1e-08

For non-fish (MF142)

Less than 200 thousand tonnes
Lower values around 10 tonnes

name	items	passes	fails	nNA	error	warning	expression
Undir.100kt	33	33	0	0	FALSE	FALSE	gildi - 2e+05 <= 1e-08
Yfir.þúsund.tonnum	33	27	6	0	FALSE	FALSE	gildi - 10000 >= -1e-08

Data for the assembly

The data from SJA09031 is better suited than the data from SJA09005

Staða á gagnasöfnum: (skref:7)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221
MF1222_lifmassi_af_beitarlondum	66	33	8504456	MF1222
MF13_FAO_timbur	24	12	22237.94, 10630.10	MF131, MF132
MF13_FAO_ext	4	1	5444, 1672	MF131, MF132
MF141_MF141_ur_sjo	1280	33	44439280, 1176859	MF141, MF142

Quality check: Data here against Eurostat fish_ld_main table.

This dataset is somewhat complex at Eurostat and it seems that near all catch from the sea is allocated to MF141. Eurostat does not have values for whaling included as far as I can tell. In addition I don’t see a way to check if there is a separation between national vessels operating in foreign waters, or bringing catch to foreign processing.

I can compare the data in the table with the values from SJA09031.Table_A.afli_landad() with following filters:

species %in% c(‘F04’, ‘F07’),
natvessr==‘IS’,
pres==‘TOTAL’,
unit==‘TPW’,
dest_use==‘TOTAL’

Compilation 2024: Update in Euorstat db API has made this validation outdated

In previous years the values from SJA09031.px were found to be nearly identical to the Eurostat published data for years that were available, so this source is used here.

Processing of SJA01110: Whaling in Iceland

A single individual (Kristján Loftsson) is responsible for all whaling in Iceland. Whaling was officially outlawed in 1988, but “Hvalur hf” was allowed a “research exception” to continue whaling some species. Kristján then decided to continue catching “langreyður” in 2022 after the company was barred from importing Residual fuel (which they need to run the steam-powered vessels) in 2018.

The statistics on whaling was published in 2023 on the px-web at Hagstofan with data reaching back to 2000.

Processing here is similar to other px tables

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Hvalveiðar og útflutningur hvalaafurða 2000-2023: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The table contains the number of animals caught (not the weight), but the mass of the exported prodcut, and the export value (kr) is also reported in the same table. There can be a significant delay between when the animal products are exported and when they are caught, and it is easy to miss the toll number for whale products (a few years whale products were exported using the same CN as pidgeon meat, but an Icelandic designation that clearly indicated the export of products from “other” ocean mamals)

In order to calculate the life-weight of the animals I will use the average weight of the species indicated in the table.

The processing uses the following service functions

SJA01110.Table_A.check_keys: Checking keys
SJA01110.Table_A.hvalveidar: Collect data

px table from local drive

content SJA01110.px
value.Ár	value.Tegund	value.value
2006 : 4	Hrefna :16	Min. : 0.0
2009 : 4	Útfluttar hvalaafurðir (tonn) :13	1st Qu.: 38.5
2010 : 4	Útfluttar hvalaafurðir (1000 kr):13	Median : 145.0
2013 : 4	Langreyður : 9	Mean : 275179.4
2014 : 4	Búrhvalur : 0	3rd Qu.: 2070.5
2015 : 4	Hnúfubakur : 0	Max. :2770851.0
(Other):27	(Other) : 0	NA

Check mapping

Whale type - mapped: 8
Whale type - not mapped: 0

Whale types missing mapping in mfa.gagnalindir_mfa_varpanir
ID	lind_cd	vinnsla_cd	lind_dalkur	eining_inn	hlutur	table_cd	mfa_cd	vara_cd	lind_texti	athugasemdir	vinnsluadferd_cd

Here I use “hlutur” as a conversion factor between tonns per animal, the mfa_cd value is used to map the actual value to MFA category

Data from the table

##    vara_cd            table_cd            mfa_cd           lind_texti       
##  Length:25          Length:25          Length:25          Length:25         
##  Class :character   Class :character   Class :character   Class :character  
##  Mode  :character   Mode  :character   Mode  :character   Mode  :character  
##                                                                             
##                                                                             
##                                                                             
##   eining_inn            hlutur            ar           gildi     
##  Length:25          Min.   : 4.50   Min.   :2003   Min.   :  27  
##  Class :character   1st Qu.: 4.50   1st Qu.:2008   1st Qu.: 171  
##  Mode  :character   Median : 4.50   Median :2012   Median : 261  
##                     Mean   :20.16   Mean   :2012   Mean   :2082  
##                     3rd Qu.:48.00   3rd Qu.:2015   3rd Qu.:6000  
##                     Max.   :48.00   Max.   :2023   Max.   :7440  
##    vinnsla           reiknad_dags       
##  Length:25          Min.   :2024-04-09  
##  Class :character   1st Qu.:2024-04-09  
##  Mode  :character   Median :2024-04-09  
##                     Mean   :2024-04-09  
##                     3rd Qu.:2024-04-09  
##                     Max.   :2024-04-09

Gögn úr SJA09005.Table_A.afli_ur_sjo()

Sanity check of the values

Doing a sanity check on whaling is either an oxymoron or a foregone failure.

Data for the assembly

The data from SJA09031 is better suited than the data from SJA09005

Staða á gagnasöfnum: (skref:8)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221
MF1222_lifmassi_af_beitarlondum	66	33	8504456	MF1222
MF13_FAO_timbur	24	12	22237.94, 10630.10	MF131, MF132
MF13_FAO_ext	4	1	5444, 1672	MF131, MF132
MF141_MF141_ur_sjo	1280	33	44439280, 1176859	MF141, MF142
MF142_hvalveidar	25	18	52047	MF142

Catch from lakes and rivers and animal hunting

The datasource here is slighlty more complex than simple fishing statistics, since many of the most prestigeous river fishing requires you to practice catch and release of all native species.

The only native land animals in Iceland are the Artic Fox, walrusses (protected) and seals. The reindeer herd in Iceland is semi-wild, and hunting licenses are sold to keep the herd in check. Other hunted species are ducks, geese and ptarmigans

PX table LAN10302 tallies salmon caught, including catch-and-release salmons.

Table LAN10301 has the fishing statistics for the primary salmon rivers, including the weight and average wegith of the catch. This allows us to estimate the total weight of salmons

Processing of LAN10302: - Salmon fishing by fishing gear used

Processing here is similar to other px tables

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Laxveiði eftir veiðiaðferð 1974-2022: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The processing uses the following service functions

SJA10302.Table_A.check_keys: Checking keys
SJA10302.Table_A.afli_ur_votnum: Collect data

px tafla:

Summary of LAN10302.px
value.Veiði	value.Ár	value.value
Fjöldi laxa alls, þ.m.t. sleppt:49	1996 : 6	Min. : 375
Netaveiði :49	1997 : 6	1st Qu.: 13421
Stangveiði :49	1998 : 6	Median : 33005
Landaðir laxar :49	1999 : 6	Mean : 38925
Stangveiði, þar af sleppt :27	2000 : 6	3rd Qu.: 51977
Hafbeit :27	1974 : 5	Max. :217649
NA	(Other):215	NA

Skoða lykla

Fish type: Mapped: 6
Fish type: Not mapped: 0

Here I am most interested in the total number of salmons, rather than splitting the salmon by fishing method. .

Data from the table

Gögn úr SJA10302.Table_A.afli_ur_votnum()

The numbers here are number of salmon (not the weight)

Processing of LAN10301.px: Average weight of salmon

I can find the average weight of a salmon in table LAN10301

Processing here is similar to other px tables

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Laxar veiddir á stöng eftir landsvæðum og völdum laxveiðiám 1998-2022: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Return the connected keys (ommit keys that are not matched)

The processing uses the following service functions

SJA10301.Table_A.check_keys(): Check mapping
SJA10301.Table_A.medalthungi_laxa: Collects the data and calculates the weighted averege of salmons caught

px table

summary of LAN10301.px
value.Atriði	value.Ár	value.Laxveiðiá	value.value
Fjöldi laxa :1324	2008 : 265	Suðvesturland : 125	Min. : 1
Fjöldi laxa sleppt:1249	2014 : 265	Laxá í Kjós : 125	1st Qu.: 16
Afli, fjöldi :1317	2015 : 265	Vesturland : 125	Median : 299
Þyngd, kg :1317	2019 : 265	Laxá í Leirársveit: 125	Mean : 1312
Meðalþyngd, kg :1324	2004 : 264	Hvítá (Vesturland): 125	3rd Qu.: 936
NA	2007 : 264	Grímsá og Tunguá : 125	Max. :70025
NA	(Other):4943	(Other) :5781	NA

Check mapping

Unit of measure: Mapped: 5
Unit of measure: Not Mapped: 0

Land/river: Mapped: 53
Land/river: Not Mapped: 0

Data from the table

The service function SJA10301.Table_A.medalthungi_laxa() does the following:

Gets the data from the table
Fetches the mapping
Calculate the total mass and the total number of salmons
Calculates the average mass per salmon
Returns the talbe

Meðalþungi laxa (yfir allar ár) eftir árum

Conversion of number of salmons caught to weight

This process is stored in the service function Table_A.lifmassi_ur_votnum() which uses the service functions above to calculate the total mass of salmon from rivers and lakes. The function uses LVCF to fill in gaps in the data.

Gögn úr Table_A.lifmassi_ur_votnum()

Data for the assembly

This data belongs to MF141

Staða á gagnasöfnum: (skref:9)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221
MF1222_lifmassi_af_beitarlondum	66	33	8504456	MF1222
MF13_FAO_timbur	24	12	22237.94, 10630.10	MF131, MF132
MF13_FAO_ext	4	1	5444, 1672	MF131, MF132
MF141_MF141_ur_sjo	1280	33	44439280, 1176859	MF141, MF142
MF142_hvalveidar	25	18	52047	MF142
MF141_afli_ur_votnum	33	33	4043.918	MF141

More data for MF142 All other aquatic animals and plants

No data exists on number of seal caught and the export of seal skins from Iceland is trivial
Hagstofan has statistics on production of salmon and other fish in fish farming. These animals should not be counted here. Farmed fish is fed either imported feed, or feed processed in fish smelting factories.

MF143 Hunting and gathering

The data for this category primarily comes from table LAN10303

The table contains a summary of the hunting of birds, reindeer, mink and other animals that can be considered pests. All these animals fall under category MF143: Hunting and gathering. The table, however, also has animals that are not food, and are discarded without becoming a part of the economy and are directly discarded (to DPO). The table also only takes into account the number of animals, so a conversion of average weigth per animal is used.

Processing LAN10303

: Hunting

Processing here is similar to other px tables

Collect values from the table: px taflan fannst
Check if the same table exists on the web-db: Dýra- og fuglaveiðar eftir tegundum 1995-2022: (The name of the table found on the Icelandic Px-web, otherwise the value is the web-error code)
Collect keys from the mfa.gagnalindir_mfa_varpanir table and match the keys to the px-table title
Use “Hlutur” as a conversion factor between number and weight (tonnes)
Return the connected keys (ommit keys that are not matched)

The processing uses the following service functions

LAN10303.Table_A.check_keys: Checking keys
SJA10303.pxweb.Table_A.veidi_a_landi: Collect data and converts number to weight

px table:

Summary of LAN10303.px
value.Ár	value.Tegund	value.value
1995 : 32	Hreindýr: 27	Min. : 1
1996 : 32	Minkur : 27	1st Qu.: 866
1997 : 32	Refur : 27	Median : 2478
1998 : 32	Blesgæs : 27	Mean : 10951
1999 : 32	Grágæs : 27	3rd Qu.: 9346
2000 : 32	Heiðagæs: 27	Max. :232936
(Other):672	(Other) :702	NA

Check mapping

Animal type: Mapped: 32
Animal type: Not Mapped: 0

Here animals such as rats, seaguls and minks probably constitute very few tonnes per year, but these animals should probably not be considered to be “entering the economy” since they are mainly being killed and discarded.

Data from the table

Gögn úr LAN10303.Table_A.veidi_a_landi()

There appears to be a break in 2004.

Data for the assembly

Staða á gagnasöfnum: (skref:10)
Vinnslueining	Fjöldi lína	Fjöldi ára	gildi	mfa_cds
MF11_landbunadur_magntolur	419	34	219688, 338791, 208, 148854, 3005	MF1221, MF112 , MF117 , MF111 , MF116
MF1211_landbunadur_stra	40	32	219688	MF1211
MF1221_landbunadur_rummal	94	33	13549008	MF1221
MF1222_lifmassi_af_beitarlondum	66	33	8504456	MF1222
MF13_FAO_timbur	24	12	22237.94, 10630.10	MF131, MF132
MF13_FAO_ext	4	1	5444, 1672	MF131, MF132
MF141_MF141_ur_sjo	1280	33	44439280, 1176859	MF141, MF142
MF142_hvalveidar	25	18	52047	MF142
MF141_afli_ur_votnum	33	33	4043.918	MF141
MF143_vilt_dyr	864	27	12946.72	MF143

Combination of data and comparison to preivous submission

Last year’s submission (is_current) is collected for comparison with the data calculated here

## Warning: Removed 1 rows containing missing values (`geom_text()`).

## MFA-Table_A.mf14.rmd done

Figures I’d like to get

Harvest of seaweed
Plant cutoffs and trimmins from municipalities - question if this should come from waste-data

MF.2 Metal ores (gross ores)

No metal ores are extracted in Iceland

MF.3 Non-metallic minerals

##  [1] "MF3"   "MF31"  "MF310" "MF32"  "MF33"  "MF34"  "MF35"  "MF36"  "MF37" 
## [10] "MF38"  "MF39"

Data that we look for is:

MF.3 Non-metalic minerals
- MF.3.1 Marble, granite, sandstone, porphyry, basalt, other ornamental or building stone (excluding slate)
- MF.3.2 Chalk and dolomite
- MF.3.3 Slate
- MF.3.4 Chemical and fertilizer minerals
- MF.3.5 Salt
- MF.3.6 Limestone and gypsum
- MF.3.7 Clays and kaolin
- MF.3.8 Sand and gravel
- MF.3.9 Other non-metallic minerals n.e.c
- MF.3.A Excavated earthen materials (including soil), only if used (optional reporting)

This information is not as readily available as many others, since quarrying and mining for non-metals is often done in small gravel pits by contractors as a part of a larger construction project. Contractors may also have their own estimates for how much they need from each mine, but the exact quantity is rarely registered with any accuracy. This means that the error in quantity can easily be several dumptrucks of earthen material, which may not be a large error in terms of financial transactions (dirt is cheap), but this becomes a major problem when the weight of earthen material is to be used as a component in the material consumption of the economy. The weight of all produce that is domestically extracted is probably compatible to the weight of gravel in these ten “error” trucks. This is what I think of as “The concrete and carrots problem”; The domestic extraction of the economy is dominated by the least accurate component in the data collection step.

Production of material from gravel pits and mines

List of gravel mines in Iceland is available online at namur.is. This site has a detailed map-view of all the registered gravel pits in iceland, the status of the mine (open/closed/environmental recovery completed), but absolutely nothing about how much material was extracted, or what purpose was for the gravel mine.

Post 2012, mine operators needed to turn in environmental impact document where they specify the volume of material that is to be removed and the duration of the mining operation. The documents on the site are, unfortunately pdf format, or CAD, and are planning documentation rather than operational summary after the fact.

Special regulations have been in effect for material mining from the ocean floor (93/1990) which specify that “after the fact” reporting must be turned to Orkustofnun

Material consumptiom based on construction activity (cement consumption)

Majority of houses are constructed as steel-reinforced poured concrete construction. The sand/gravel requirement for this can easily be estimated using the Eurostat Sand and Gravel model.

\[ \textrm{Sand and gravel input (t)} = \textrm{cement consumption (t)}\times 6.09 \]

The total cement consumption is almost certainly equal to the import of cement + national production - cement export. This data is relatively accessible; cement import is available on the trade-database, and the manufacturing of cement in iceland is well documented (ended in 2012). Her

Imports of Cement

## Sæki gögn af diski frá: C:/Users/ThorsteinnA/OneDrive - Public Administration/Documents/R/EW-MFA//_GognInn/sementsinnflutningur.parquet

The dataset here is:

Toll numbers: 25232100, 25239000, 25232900, 25233000, 25231000
Number of importers: 302
Average import: 7.7165^{4} tonn

Native production of cement

This data was obtained through personal communication with expert at Sementsverksmiðjan. This is a limited data that is not going to change since local manufacturing ended in 2012.

Consumption of cement and consumption of sand and gravel (MF.3.8) in concrete manufacture

Here I have the data needed to calculate the needed information.

Data for the assembly

##                                 Length Class  Mode
## MF11_landbunadur_magntolur      10     tbl_df list
## MF1211_landbunadur_stra         10     tbl_df list
## MF1221_landbunadur_rummal       10     tbl_df list
## MF1222_lifmassi_af_beitarlondum 10     tbl_df list
## MF13_FAO_timbur                 10     tbl_df list
## MF13_FAO_ext                    10     tbl_df list
## MF141_MF141_ur_sjo              10     tbl_df list
## MF142_hvalveidar                10     tbl_df list
## MF141_afli_ur_votnum            10     tbl_df list
## MF143_vilt_dyr                  10     tbl_df list
## MF38_steypa                     10     tbl_df list

Gögn úr Table_A.MF3.jardefni_vegagerdin()

Here we get a spike in cement construction around the year 2008 during the construction of Kárahnjúkavirkjun, the largest hydroelectric dam in Europe.

Amount of earthen material used in road construction

A major revision of methodology and datasources was done in 2024. In previous years I used data collected from the environmental reports from the national road authority Vegagerðin. This environmental report was canceled in 2018, and I had to ask specifically for data from the report author. In the report they summarized material consumed (volume) used for the foundation/leveling part of the road construction, the volume of carring layers and the volume of surface layers. The data in this environmental report did not contain how much material was removed from the path of the road, but this material is commonly used to build up and level the new road. This material is rarely fully removed from the road site, but commonly used used in the underlayers for the road, and not transported more than 50-100m. The removed material can alternatively be spread out round the construction site. My understanding of the compilation manual is that this short distance transport of material should therefore not be included in the actual material consumption for the road.

In modern reconstruction of roads the material from the previous road is commonly used for the new road and the old road site restored to close to natural state. By only counting the material added to the new road and not subtracting the removed material, I am likely to double-count material consumption for the road as the old road is no longer in existance.

Estimate from the length of roads in Iceland

Eurorstat provides me with basic estimates (using German coefficients) for material used in road construction based on the length of roads and length of new roads constructed. Statistics Iceland has a published table on national roads in Iceland. This data is updated regularly (and can be extended). The data has a peculiarity that classification of roads can change between years, without new construction being done. The national roads can therefor be “shortened” between years, which may be due to re-classification, or closing of roads. This data is collected from the map-data from Vegagerðin hosted by Landmælingar Íslands.

Road construction in Iceland is also very different from road construction in Germany. Roads in Iceland are built across wetlands, deserts, soft lavafields or across floodplains. The construction on rural roads relies on re-construction of the foundation for the road, which sometimes sits on a several meter tall berm with sloped sides. The common knowledge here is that road construction requires more quarrying than in most countries, with the exception perhaps of Finland. Iceland also has many small, but complex harbors often with retaining walls and reinforcement of the coastline against the weathering of the ocean.

I am therefore very hesitant in using the Eurostat coefficients to estimate the quarry material needed for this type of infrastructure projects

Amount of material used in road-construction contracts

The road authority (Vegagerðin) manages all national roads in Iceland, and is also the financing manager of most major road constructions in municipal areas. Road constructions at Vegagerðin are put up for public bidding, either nationally or on the EU/EFTA area. The construction auctions are advertised online, through the site of “Ríkiskaup” and in a publication called Framkvæmdafréttir, which Vegagerðin publishes. This publication also contains information on successful bids for construction and in some cases articles that describe the concluded projects.. Publications from 1993 can be found at the timarit.is website, but articles from 2002 are on the above link.

The PDF files from the publication are unfortunately not well suited for scraping, since the data tables in the publication is somewhat heterogeneous, and multiple layouts of the magazine have bee attempted. The data is also not available on table form from Vegagerðin (discussion with people there) directly, so the only available data gathering is direct entry from reviewing the PDF documents. Here I am most intersted in the “successful bid” documentation, as this summarizes a contract that was signed between Vegagerðin and a contractor that will take on the construction. The information/columns collected here are the following

heimild: The citation includes the issue, volume and publication year
utbod_nr: The serial number for the bidding
vegur: Title/Description of the road and project. Recent data includes the road number, which is connected to the location of the road.
dags_utbod: The date when the bid was posted
dags_verklok: The date for when the project is supposed to be finished. This is here used as the “delivery date” of all material for the construction, although there may be delays on the contract, or the construction may take several months to a year. The projects are, however, often partitioned down to short road segments that are completed with few months. Vegagerðin only pays 20% of estimated cost submitted in the winning bid before the construction starts, and the rest is paid after the project has been completed.
vegalengd_km: The distance of the road being constructed (if applicable)
skeringar_m3: The cubic meters of material that is dug-out from the path of the road. This material may or may not be used for the road construction, but may be used as filling or foundation for the road.
bergskeringar_m3: The cubic meters of boulders, hard-rock and cliffs that need to be removed (using dynamite or freeze-cracking) from the path of the road. This quantity is sometimes included in the “skeringar_m3” value when the hard-rock volume is small. Blasted rock is nearly allways used for construction of the road after milling.
fyllingar_m3: Volume of earthen material used for leveling the road-bed. Modern road construction (post 1980) have substantial volume of this fill-up as the roads are built for higher traveling speeds. This filling material is commonly taken from the “skeringar_m3” volume (e.g. by using a bulldozer to push the earth into piles).
Fláafleigar_m3: Volume of material used to build side-berms on the roads. This side berm is there to protect the foundation of the road from erosion and provide a safety zone for drivers that miss the road. This material is almost always taken from the “skeringar_m3” part in the construction. This volume is only included in the bidding if the volume of the fláafleigar part is substantial, otherwise it is reported in “fyllingar og fláafleigar”.
Burdarlag_m3: Volume of the carrying layer of the road. The grain size of this layer is commonly greater than 63mm in order to provide a structural cushion and a drainage layer for the road. This layer must be from “bergskeringar_m3” type of material (i.e. not sand, dirt or pebbles). It can be assumed that this material comes from quarrying mines unless the project has specific “bergskeringar_m3” volume.
slitlag_m3: Gravel layer that fills into the pores of the “burðarlag_m3” and provides a “smooth” layer which can be paved over. This layer has particulate size 0/30 mm. The “Slitlag_m3” is the final road layer for old style gravel roads.
klaedning_m2: Square meters of smoothing layer on the road. This layer is commonly 1-2 cm thick from rock measuring approximately 10mm or smaller. This layer contains some asphalt binding agent. These layers became common in post 1990 construction. Most roads through the Icelandic coutry side are coated with this type of surface (approximately 85% of roads in 2023). Rocks for this layer are tightly controlled in size and must come from a gravel quarry. The downside of this type of covering is high road noise, less resistance against wear and loose material that can be generated in freeze-thaw cycles.
malbik_m2: Square meters of asphalt layers in the construction. These layers are commonly 3-5 cm thick and contain higher fraction of bitumen and asphalt than “klaedning” layers. The price of asphalt construction is much higher than “klaedning”, but the resulting road is smoother, more wear-resistant layer that cars can travel over more quietly. This construction is therfore preferred type of surface for heavily traveled roads and roads in populated areas
fraesing_m2: In recent years contractors have been using road-graders to remove asphalt layers from the road before they are re-paved. The graded material is re-coated and used in “klaedning”. We can therefore use this “fraesing_m2” material (assuming that it is 1-2 cm thick) to subtract material consumption for slitlag construction
fragangur_flaa_m2: The surface area of road shoulders that are coated with dirt or grass. Here I will assume that this surface area does not require material in the construction
steypa_m3: Cubic meters of concrete in the project. The sand and gravel demand of concrete struction is calculated from cement import and this information is not used
stal_tonn: Tonnes of steel in the construction. This information is not used here.
Tegund verks: The type of project
kr: The winning bid cost in krónur.
Verktaki: The contractor that won the bid. This information can be used to compare to income reporting of the same contractor (from tax-record database)

The current data is

Number of rows: 1502
First year in data: 1993
Newest year in data: 2024
Number of years (there may be empty years): 32
Number of different types of projects: 24, or: Smíði stálbita, Efnisvinnsla, Ferjurekstur, Nýbygging, Brúarsmíði, Vegsmíði-Þéttbýli, Skiltasmíði, Annað, Húsasmíði, Landgræðsla, Vegmálun, Viðgerð-styrking, Varnargarður, Malbikun, Ræsi, Endurbygging, Rannsóknir, Málmsmíði, Lokað útboð, Eftirlit, Mölburður, Breikkun, Hönnun, Jarðgöng. Some of those projects include no consumption of material

Example data from 2018 for few key-types o

Sample projects from 2018
name	Malbikun	Endurbygging	Vegsmíði-Þéttbýli	Varnargarður	Viðgerð-styrking	Nýbygging
heimild	4.tbl. 26 árg. Nr 683	4.tbl. 26 árg. Nr 683	4.tbl. 26 árg. Nr 683	5.tbl. 26 árg. Nr 684	6.tbl. 26 árg. Nr 685	7.tbl. 26 árg. Nr 686
dags_verklok	2018-09-01 00:00:00	2019-08-01 00:00:00	2018-09-15 00:00:00	2018-11-01 00:00:00	2019-11-01 00:00:00	2019-10-01 00:00:00
vegalengd_km	18	8.3	0.85	NA	0.92	6.1
skeringar_m3	NA	17000	1000	NA	28000	36300
bergskeringar_m3	NA	NA	NA	NA	NA	700
fyllingar_m3	NA	10000	12500	3500	20100	81000
flaafleigar_m3	NA	5500	NA	6500	NA	33000
styrktarlag_m3	NA	10800	5900	NA	4750	41000
burdarlag_m3	NA	18700	2100	NA	1700	11000
slitlag_m3	NA	NA	NA	NA	6000	NA
klaedning_m2	NA	66500	9150	NA	NA	47000
malbik_m2	125464	NA	NA	NA	NA	NA
fræsing_m2	NA	NA	NA	NA	NA	NA
fragangur_flaa_m2	NA	79000	13300	NA	NA	NA
steypa_m3	NA	NA	NA	NA	NA	NA
stal_tonn	NA	NA	NA	NA	NA	NA
kr	418472075	488297265	118893244	47667600	85733000	297872980

The table above shows a few sample projects from 2018, few of the information in the table have been ommited for clarity. The collected data was initially stored in a PMDB file, but is later migrated to the UmhverfisTolfraedi database. Maintenance of this data is managable if the maintenance is done each year, but initial data entry was quite substantial.

Range of construction projects: 1993 to 2026 to
Total distance of road construction for some of the main project types
- Reconstruction: 1438.86km
- New construction: 1298.976km
- Area of re-surfacing: 5.3268314^{7}m2
Types of constructions: Smíði stálbita, Efnisvinnsla, Ferjurekstur, Nýbygging, Brúarsmíði, Vegsmíði-Þéttbýli, Skiltasmíði, Annað, Húsasmíði, Landgræðsla, Vegmálun, Viðgerð-styrking, Varnargarður, Malbikun, Ræsi, Endurbygging, Rannsóknir, Málmsmíði, Lokað útboð, Eftirlit, Mölburður, Breikkun, Hönnun, Jarðgöng

Some of those project do not include any information on gravel quantity e.g. Vegmálun = road painting, Landgræðsla = re-planting, Eftirlit = surveying. Other projects, such as Varnargarður is for construction of berms, shorelines etc. The material consumption in these projects is somewhat unique and needs to be considered separately from other projects.

The road construction projects are here separated into three phases

The foundation phase. This phase can require large volume of gravel and foundation rocks (basalt) in order to flatten and straighten the road. In this phase, material from the surrounding can be used.
The road strengthening layer phase. This phase requires coarse and strong earthen material to be used in order to create a foundation for the road that allows water to seap through, but stops the road from moving around when traffic travels on the road.
The surface finishing phase. This layer is fine ground gravel, sometimes bound with tar, or oil layer. Asphalt is only used for roads where road noise must be reduced, or where traffic is medium to heavy. Cement bound roads were first built in 2014, and are still relatively rare.

The data contains dates for when the offer was announced and when the project is to be delivered. Some projects are only several months, whereas other projects last more than a year. Multi-year projects are rare. It is, however not impossible to create a spread-function to allocate material consumption down to months, but this type of manipulation would reduce the year-to-year fluctuations in material consumption somewhat.

Estimating the amount of earthen material for road foundation phase

Up until 1970, roads in Iceland were built “on top of” and around the landscape, with maximum speed limit of 35-50 km/hr. The rural roads were gravel roads that were re-surfaced by grating a new layer every two to three years. The ring-road (highway 1) was completed in 1974, with the construction of the 6 km long single lane bridge across Skeiðarársandur. The last part of the ringroad was paved with oil-bound gravel layer in 2019.

Modernization of the road system in Iceland has been focused on buiding every road on a raised layer (berm; fyllingar) and in some cases a sacrificial taperin (fláaeining) is added in order to reduce erosion and collapse of the berm. An added benefit is that drivers are less likely to die if they slip of the road as the taper layer is constructed from softer material.

The upper limit for material consumption in this part of the construction is therefore the sum of the berm and taper used, or:

\[ V^F= V(\textrm{fyllingar_m3})+V(\textrm{flaafleigar_m3}) \]

In most cases, this material is moved around using buldozers and other excavation equipment from the surroundings of the road. This removal can also be a physical dig-up of the previous road. I therefore map this volume into MF.3.A: Excavated earthen material (including soil), only if used (optional reporting)

Reclamation of the previous roads became part of the law in 2014. The volume of this removal (skering) is included in the data in two parameters: skering=gravel cuts and bergskering = clif-cuts:

\[ V^S = V(\textrm{skeringar_m3})+V(\textrm{bergskeringar_m3}) \]

Some of this removal volume can be used in the foundation (burðarlag) layer of the road construction (here reported as \(V(\textrm{burdarlag_m3})\)). In many cases, the Buðarlag part of the road needs to be transported in from gravel pits and mines.

The net transport of material to the road from the road-foundation is therefore:

\[ V^\circ = V^F+V(\textrm{burdarlag_m3}) - V^S \]

In rare cases (e.g. in tunneling projects), the value of \(V\) is negative, as the amount of removed material is more than what takes to build the road. In these cases the excess material is either leveled out around the road, or transported to another construction site. The best case scenario here is therefore that the berm/taper/carrying layer construction requires no transport of material, or is zero (never negative)

\[ V^\circ = \left\{\begin{array}{l c l }V^F+V(\textrm{burdarlag_m3}) - V^S \\ 0 \qquad \textrm{if} \qquad V^F+V(\textrm{burdarlag_m3}) - V^S<0\end{array}\right. \]

The figure above shows the total “filler” material as blue dots, but the effective volume (V0 is shown as red dots). Some of this material comes from basalt mines (MF.3.1), whereas some of it is earthen material/gravel (MF.3.8). The blue-dotsare the volume values that were reported prior to 2024.

Here I will use density of dry gravel as 1.680 tonnes/m3

Estimating the amount of marble, granite, sandstone and basalt

The material used for road construction is a combination of gravel and basalt boulders. I do not have a good indication of the actual ratio of basalt to gravel, but the data I have gives some indication of how much basalt is removed in the construction. If I assume that this material is all consumed in road construction, I can calculate a proxy value according to:

\[ f = \frac{V(\textrm{bergskeringar_m3})}{V(\textrm{bergskeringar_m3})+ V(\textrm{skeringar_m3})} \]

provided that \(V(\textrm{skeringar_m3})\neq 0\) for the project

Here I will use the fraction (0.1688) for basalt/(basalt + gravel) to split the \(V^\circ\) value above between MF.3.1, MF.3.8 and MF.3.A

Estimating the material for gravel road top layer

The material used for gravel layer (\(V(\textrm{slitlag_m3})\)) is here always assumed to be transported from a quarry and consist of processed basalt rock. The quarries in Iceland are first and foremost basalt mines where the rocks with correct particle size are milled out. The sand/gravel production from the mines is first and foremost used for road foundation.

Estimating material for paving and asphalt layers

The thickness of “klæðning” is assumed to be 0,015 m (15 cm). The amount of bitumen and oil binders in these layers is approximately 3% of the total volume of the coating. The material volum of this layer is therefore

\[ V^K = A(\textrm{klaedning_m2})\times 0.015 \times 0.97 \] The thickness of asphalt layer is 3-5 cm, or 0,03-0,05 m, with 10% of this volume filled with bitumen, asphalt and other oil binders. The oil binders are mainly gap-filling in the coarse rock layer (maximum volume fractions filled with close packed spheres is 64%, but apparently the asphalt reduces the packing by 10%. thus i assume 10% fill-volume of bitumen). In some projects they recycle grated material from the road. The volume consumption of asphalt layers is therefore:

\[ V^A = A(\textrm{malbik_m2})\times 0.04 \times 0.9 - A(\textrm{fræsing_m2}) \times 0.015\times 0.9 \]

Road-surfacing projects are most commonly only awarder for projects starting in march and ending in november. This makes the “spreading” function return largely the same values as the annual values. It is important to note the difference in vertical scale in this last graph. This material is classified into MF.3.1

Data for the assembly

##                                 Length Class  Mode
## MF11_landbunadur_magntolur      10     tbl_df list
## MF1211_landbunadur_stra         10     tbl_df list
## MF1221_landbunadur_rummal       10     tbl_df list
## MF1222_lifmassi_af_beitarlondum 10     tbl_df list
## MF13_FAO_timbur                 10     tbl_df list
## MF13_FAO_ext                    10     tbl_df list
## MF141_MF141_ur_sjo              10     tbl_df list
## MF142_hvalveidar                10     tbl_df list
## MF141_afli_ur_votnum            10     tbl_df list
## MF143_vilt_dyr                  10     tbl_df list
## MF38_steypa                     10     tbl_df list
## MF310_road_base                 10     tbl_df list
## MF31_road_base                  10     tbl_df list
## MF38_road_top                   10     tbl_df list
## MF38_topcoat                    10     tbl_df list
## MF31_asphalt                    10     tbl_df list

Calculated total consumption of earthen material for road construction

The total consumption of earthen material is the sum of these factors, or:

\[ V_T = V^\circ + V(\textrm{slitlag_m3})+V^K+V^A \]

The interesting value here is the \(VT\) value. The sub-parts of are stored in the assembly. The spikes around 2008 are due to heavy construction of roads around the Kárahnjúkar powerplant. The high consumption of gravel in and around the year 2000 is explained by major re-construction of the road system in Iceland, when many roads were being rebuilt from being gravel to being “levelled” roads (on berms etc).

Material consumption for harbor construction and flooding barriers

Harbor constructions and flood preventions became an important part of the Vegagerðin’s task in 2000. This includes construction of harbor walls, flood-plane berms (that are used to guide glacial rivers to common paths) and rock-slide berms. The material used for this construction is a combination of basalt bolders and gravel fillers. These projects are generally only done over the summertime. The “varnargarður” projects also include construction of protective barriers on floodplanes and protective structures to divert lavaflows.

In this data “skeringar_m3” value is used to identify how much of bolders are moved and re-constructed, fyllingar_m3 indicates the quantity of gravel, but flaafleigar_m3 is used to record the volume of bolders mined and placed. The total consumtion of basalt is therefore

\[ V_\textrm{basalt} = V(\textrm{flaafleigar_m3})-V(\textrm{skeringar_m3}) \] and the consumption of gravel is

Vegagerðin does not manage projects where cities fill in shorelines in order to gain more building space for housing. These project are often de-facto recycling sites for construction waste. The law in Iceland allows municipalities to build up to 60m wide barrier regions on shorelines to protect the nearby housing projects. New housing projects can then be developed on the new shoreline after 5 year settling time, and the shore-line can be moved another 60 meters. This practice is not covered in this statistics as the data is not available.

Data for the assembly

##                                 Length Class  Mode
## MF11_landbunadur_magntolur      10     tbl_df list
## MF1211_landbunadur_stra         10     tbl_df list
## MF1221_landbunadur_rummal       10     tbl_df list
## MF1222_lifmassi_af_beitarlondum 10     tbl_df list
## MF13_FAO_timbur                 10     tbl_df list
## MF13_FAO_ext                    10     tbl_df list
## MF141_MF141_ur_sjo              10     tbl_df list
## MF142_hvalveidar                10     tbl_df list
## MF141_afli_ur_votnum            10     tbl_df list
## MF143_vilt_dyr                  10     tbl_df list
## MF38_steypa                     10     tbl_df list
## MF310_road_base                 10     tbl_df list
## MF31_road_base                  10     tbl_df list
## MF38_road_top                   10     tbl_df list
## MF38_topcoat                    10     tbl_df list
## MF31_asphalt                    10     tbl_df list
## MF31_harbor_walls               10     tbl_df list
## MF310_berms                     10     tbl_df list

Earthen material extracted from the seafloor

This data originated from Orkustofnun, which issues lisences for seafloor mining. These projects are first and foremost dredging operations, where the material is either dumped at sea. In some cases, limestone is extracted, which is then brought to shore or deposit the material with processors. This material is first and foremost used for house constructions.

Gögn úr Table_A.MF3.jardefni_ur_sjo()

This datasource is not easily accessible, since Orkustofnun is notoriously bad at maintaining their dataset due to rapid turnover of employees. I have estimated some of the missing data using previous years and the financial filing of Björgun ehf., which is the main dredging operator in Iceland.

Data for the assembly

##                                 Length Class  Mode
## MF11_landbunadur_magntolur      10     tbl_df list
## MF1211_landbunadur_stra         10     tbl_df list
## MF1221_landbunadur_rummal       10     tbl_df list
## MF1222_lifmassi_af_beitarlondum 10     tbl_df list
## MF13_FAO_timbur                 10     tbl_df list
## MF13_FAO_ext                    10     tbl_df list
## MF141_MF141_ur_sjo              10     tbl_df list
## MF142_hvalveidar                10     tbl_df list
## MF141_afli_ur_votnum            10     tbl_df list
## MF143_vilt_dyr                  10     tbl_df list
## MF38_steypa                     10     tbl_df list
## MF310_road_base                 10     tbl_df list
## MF31_road_base                  10     tbl_df list
## MF38_road_top                   10     tbl_df list
## MF38_topcoat                    10     tbl_df list
## MF31_asphalt                    10     tbl_df list
## MF31_harbor_walls               10     tbl_df list
## MF310_berms                     10     tbl_df list
## MF30_ur_sjo                     10     tbl_df list

Model for material consumption in road construction

Eurostat provides the following coefficient for construction and maintenance of roads in Germany.

Tonn sand and gravel per km in road construction (Germany)
road type	new construction (tonn/km)	maintenance (tonn/km)
Highway	28383	518
National roads	9692	151
Federal state roads	8719	76
District roads	6777	65
Local roads	6886	81

The dataset from Vegagerðin above allows us to estimate the material consumption per km of construction for different types of construction. Here I assume that the density of all rock material is 1680 kg/m3. The volume calculations are done the same way as above. Projects that give negative volume (removal is greater than the foundation raising/foundation of the project).

The box plot is split down by types of construction and the decade of the construction, since we expect that project technology will have changed substantially over the year. The types of construction projects here are

Viðgerð-styrking: Regrading and improvinng gravel roads
Vegsmíði-Þéttbýli: Road construction in towns and cities
Nýbygging: Building of new roads
Malbikun: Paving/resurfacing of roads - most common maintenance projects, but these project are most commonly not specified to roads, but rather all areas within a district. Due to this I could not include these roads in the statistics at this point.
Endurbygging: Reconstruction of road/improvement - These proejcts often include major overhall, bridge buildings and drainage improvements
Brúarsmíði: Bridge building. The length of the bridge is commonly not included, but I approximated the length from maps
Breikkun: Road broadening.

Material consumption (tonn/km) for different types of construction, including decade
decade	name	Annað	Breikkun	Brúarsmíði	Efnisvinnsla	Endurbygging	Mölburður	Nýbygging	Ræsi	Vegsmíði-Þéttbýli	Viðgerð-styrking
1990-s	datapoints	NA	NA	1.000	1.0000	33.000	4.0000	36.00	2.00	4.000	41.0000
2000-s	datapoints	2.0000	2.000	13.000	NA	112.000	3.0000	91.00	1.00	8.000	38.0000
2010-s	datapoints	NA	3.000	2.000	NA	83.000	NA	18.00	NA	9.000	1.0000
2020-s	datapoints	NA	1.000	2.000	NA	48.000	NA	18.00	NA	8.000	NA
1990-s	stdev	NA	NA	NA	NA	12250.279	140.7045	12624.52	25214.38	3321.631	12489.4779
2000-s	stdev	561.4639	16037.182	11309.277	NA	11137.818	278.4351	11940.21	NA	12029.575	13077.5775
2010-s	stdev	NA	4878.837	14521.155	NA	7226.428	NA	12354.73	NA	10257.346	NA
2020-s	stdev	NA	NA	1284.274	NA	10211.150	NA	11563.32	NA	8036.369	NA
1990-s	tonn_km_mean	NA	NA	46666.667	144.7554	18493.620	313.7855	19884.18	27527.56	23987.563	13311.0538
2000-s	tonn_km_mean	2402.9851	18060.000	20404.290	NA	15428.701	571.1774	20354.18	40745.81	25440.470	13148.5192
2010-s	tonn_km_mean	NA	6268.166	22868.007	NA	9401.969	NA	32165.77	NA	26795.910	697.2458
2020-s	tonn_km_mean	NA	11315.069	22273.073	NA	11573.912	NA	28809.06	NA	13028.064	NA

The table above shows the mean consumption per km and the standard deviance from mean, as well as the number of underlying datapoints in the dataset. The table shows clearly that there is no statistical significance between material consumption over the decade, so I omitt this dimensions

Material consumption (tonn/km) for different types of construction, all projects
name	Annað	Breikkun	Brúarsmíði	Efnisvinnsla	Endurbygging	Mölburður	Nýbygging	Ræsi	Vegsmíði-Þéttbýli	Viðgerð-styrking
tonn_km_mean	2402.9851	11039.928	22344.70	144.7554	13312.37	424.0963	22488.39	31933.64	22236.61	13076.18
stdev	561.4639	9713.341	11852.96	NA	10541.07	233.8154	12751.38	19393.89	10855.72	12690.59
datapoints	2.0000	6.000	18.00	1.0000	276.00	7.0000	163.00	3.00	29.00	80.00

This table shows that there is little reason to separate the material consumption by types, unless we combine Brúarsmíði (bridge building) with Nýbygging (new construction), and Breikkun (widening) with Viðgerð (repair of gravel roads) and endurbygging (reconstruction) with Vegsmíði-Þéttbýli (road-building in dense areas). This combination is also somewhat logically sensible.

Mapping between different projects and broader categories
tegund_verks	project_type	tonn_km_mean
Brúarsmíði	New construction	22344.70
Nýbygging	New construction	22488.39
Vegsmíði-Þéttbýli	New construction	22236.61
Endurbygging	Reconstruction	13312.37
Breikkun	Repairs-broadening	11039.93
Viðgerð-styrking	Repairs-broadening	13076.18

Material consumption (tonnes/km) by construction type for roads in Iceland
project_type	tonn_km_mean	stdev	datapoints
New construction	22441.31	12379.69	210
Reconstruction	13312.37	10541.07	276
Repairs-broadening	12934.11	12470.16	86

The values in this table fall in-between the new-construction values for new construction of Highways and national roads for the road system in Germany that Eurostat provides. This is an interesting observation, but for this statistical project, we don’t need to use this estimate (except for internal validation), since the data I have access to is for more than >90% of all the road construction projects in Iceland.

Combination of data and comparison to preivous submission

Last year’s submission (is_current) is collected for comparison with the data calculated here

## Warning: Removed 8 rows containing missing values (`geom_text()`).

## MFA-Table_A.biti.mf30.rmd done

MF.4 Fossil energy materials/carriers

No coal or oil extraction in Iceland

Build-output

The final step in this process is to gather all the data and store it in the work-database. The data in the assembly is stored after aggregating the data so that it fits into the EW-MFA questionnaire. The data assembly has the data on highest detail, which I can use to troubleshoot the data processing. The disaggregated

Lykil lýsing á gagnaramma
nafn	dalkar	radir	magn
MF11_landbunadur_magntolur	10	419	7.105460e+05
MF1211_landbunadur_stra	10	40	2.196880e+05
MF1221_landbunadur_rummal	10	94	1.354901e+07
MF1222_lifmassi_af_beitarlondum	10	66	8.504456e+06
MF13_FAO_timbur	10	24	3.286804e+04
MF13_FAO_ext	10	4	7.116000e+03
MF141_MF141_ur_sjo	10	1280	4.561614e+07
MF142_hvalveidar	10	25	5.204700e+04
MF141_afli_ur_votnum	10	33	4.043918e+03
MF143_vilt_dyr	10	864	1.294672e+04
MF38_steypa	10	34	2.958536e+07
MF310_road_base	10	33	4.940889e+07
MF31_road_base	10	33	1.003606e+07
MF38_road_top	10	33	3.228294e+06
MF38_topcoat	10	33	1.792576e+06
MF31_asphalt	10	33	3.577069e+05
MF31_harbor_walls	10	25	1.188965e+06
MF310_berms	10	25	3.240801e+06
MF30_ur_sjo	10	72	2.745673e+07
Samtals	10	10	1.950043e+08

Storage on database

Here I store the data in a tmp table and then run sql queries in the database.

truncate table mfa.gogn_tafla_a


insert into mfa.gogn_tafla_a(
    table_cd, mfa_cd,  ID_gagnalind_mfa_vorpun,
    ar, gildi, eining_inn, vara_cd,
    lind_texti, vinnsla, reiknad_dags)
select 
    table_cd, mfa_cd, ID_gagnalind_mfa_vorpun,
    ar, gildi, eining_inn, vara_cd,
    lind_texti, vinnsla, reiknad_dags
from tmp_gogn_tafla_a;

drop table tmp_gogn_tafla_a;

EW-MFA Table_A: Processing script

Þorsteinn Aðalsteinsson

2024-03-19

Versions

What is the project

MF.1 Biomass

MF11 Crops:

Processing of harvest figures (LAN10103.px)

Px table stat:

Key check

Data from the px-table:

Data to assembly

Combination of data and comparison to preivous submission

MF12: Crop residues (used)

MF121 Crop residues

MF1211 Straw

Data to assembly

MF1212 Other crop residues (sugar and fodder beats, beat leaves etc.)

MF122 Fodder crop and grazed biomass

MF1221 Fodder crops (including biomass harvest from grassland)

Data to assembly

Quality check of data against values in apro_cpnh1

MF1222 Grazed biomass

Processing and mapping for number of animals alive in Nomvember (LAN10102.px)

the px table summary

Key check

Data from the px file

Processing of th enumber of slaughtered animals (LAN10205.px)

px table summary

Mapping check

Processing and data outcome

Processing of pre 2012 figures (LAN10201.px)

px table summary

Mapping check

Processing of values from px table

Combination and final processing:

Data for the assembly

Combination of data and comparison to preivous submission

MF13 Wood

MF13.1 Timber (Industrial roundwood)

MF.1.3.2 Wood fuel and other extraction

processing

Information from FAO

Extending the FAO data

Combination of data and comparison to preivous submission

MF14: Wild fish catch, aquatic plants and animals, hunting and gathering

MF141 Wild fish

Processing of SJA09005, Catch on sea

px table from local drive

Check mapping

Data from the table

Sanity check of the values

Values for the collection

Processing of: SJA09031; Catch and values by species, type of landing and months:

px tafla

Check mapping

The values

Validate check on the data

Data for the assembly

Quality check: Data here against Eurostat fish_ld_main table.

Processing of SJA01110: Whaling in Iceland

px table from local drive

Check mapping

Data from the table

Sanity check of the values

Data for the assembly

Catch from lakes and rivers and animal hunting

Processing of LAN10302: - Salmon fishing by fishing gear used

px tafla:

Skoða lykla

Data from the table

Processing of LAN10301.px: Average weight of salmon

px table

Check mapping

Data from the table

Conversion of number of salmons caught to weight

Data for the assembly

More data for MF142 All other aquatic animals and plants

MF143 Hunting and gathering

Processing LAN10303