Historia: History: |
Source data was separate mineral deposit databases (Au, Zn, Ni, PGE, U, Cu, Industrial miner-als, FODD, old deposit database) which were combined into Mineral Deposit Database. Data-bases didn’t contain detailed production data, this data was imported from excel-sheets compiled by K. Puustinen (http://weppi.gtk.fi/aineistot/kaivosteollisuus/). Original data was based on all public information on the deposits available including published literature, archive reports, press releases, company Internet pages, and interviews of exploration geologists. In future the data sources will be the same and database will be updated whenever new data is available or new deposit is found. Deposits in the source databases were included based on certain criteria depending on the data-base. Such criteria are not in use anymore, every promising prospect will be inserted to the data-base. The original criteria were following: Au: A prospect must contain at least 1 ppm Au for a 1 m section or at least 0.5 ppm Au for a 5 m section. Zn: A prospect must have at least one drill hole with a grade of >1% Zn for >1 m or >0.5 % Zn for >5 m. Ni: A prospect must have at least one drill hole or continuous outcrop sample with a grade of >0.5 % Ni equivalent >1 m. PGE: The threshold for a prospect to be included into the database varies geographically. Cu: A prospect must contain at least 1.5% Cu for a 1 m section or 20 m at 0.5% Cu, or at least 0.2 Mt ore at 0.2% Cu. FODD: Deposits with the following metals are included in the FODD: Ag, Au, Be, Co, Cr, Cu, Fe, Li, Mn, Mo, Nb, Ni, Pb, Pd, Pt, Rh, REE, Sc, Sn, Ta, Ti, U, V, W, Y, Zn and Zr. Only such deposits where one or several of these metals form the majority of the value of the case are in-cluded, and only when there is a resource estimate of some kind in the primary reports on the deposit. U, Industrial minerals and Ore deposit database: Unknown criteria. |
Prosessointihistoria: Process step: |
Abstract The source data for the Mineral Deposit Database were separate commodity-based databases (Au, Zn, Ni, PGE, U, Cu, Industrial minerals, FODD, old Ore deposit database). Separate mineral deposit databases were combined together and data was processed in MS Access and Excel to fit the new database structure. Old databases contained a lot of descriptive text fields, so the process was done mainly manually. Some of the deposits were found in several different databases, mainly Ni+PGE, Au+Cu and Au+U. Information of these deposits was combined to form a single deposit.
Detailed description of the data processing
Ni, PGE, Cu, U and IndMin databases Ni, PGE, Cu, U and IndMin databases had similar database structures, only the amount of tables varied. PGE database had most tables, so it was chosen as a primary database where other databases were combined. Before combination, integrity of data was checked and repaired from the original databases. New Deposit_IDs were created since every database had a running number as Deposit_ID. IDs were formed by adding the original Deposit_ID to a number given for each database: PGE= 1000 + Deposit_ID Ni= 2000 + Deposit_ID Cu= 3000 + Deposit_ID U= 4000 + Deposit_ID IndMin= 5000 + Deposit_ID IndMin database had a little different structure than other databases; Minerals table was related to Host Rocks table, not to main table (Deposit) as in other databases. Uranium database had few fields which were not in the chosen primary database (PGE). These fields were imported as separate table to the combination database. Amount of deposits in combination database were PGE:35, Ni:80, Cu:36, U:67 and IndMin:96.
Au and Zn databases Au and Zn databases were structurally different than Ni, PGE, Cu, U and IndMin databases. These two databases were first combined together. Integrity of data was checked and repaired from original databases before combination. There was one deposit in Au database which was not in the main table but data was found in all other tables (Mäkrä). This was added to the database. New Deposit_IDs were created by adding the original Deposit_ID to a number given for each database: Au= 6000 + Deposit_ID Zn= 7000 + Deposit_ID In the original databases links to the figures were inserted to several multiple columns in one table, e.g. Outcrop_photo1, Outcrop_photo2, Outcrop_photo3.... These multiple columns had to be processed to normal database format (one-to-many relationship). All references were listed in one memo-style field. Every row had to be separated to form one record with Deposit_ID and Reference_ID. Database was modified to fit to the Ni+PGE+Cu+U+IndMin -database structure. There were some fields which was not in the Ni+PGE+Cu+U+IndMin -database. These were separated to different tables. Amount of deposits in combination database were Au:209 and Zn:117.
Combining Ni+PGE+Cu+U+IndMin and Au+Zn databases ->Migration database The comparison between fields and information within them was made between two combined databases. After that Au+Zn -database was imported to Ni+PGE+Cu+U+IndMin -database with Append-queries. Combination did produce some fields which belonged only to one dataset and the end result at this point was quite messy. There were several duplicate deposits in the Migration database and these had to be found. Some of them were found simply by the same name or same coordinates, but there were still many deposits which had different names and slightly different coordinates in source databases, for example Vihanti and Vihanti-U. Deposits were compared by alternative names and some duplicates were found. Final checking was done spatially with few hundred meters buffers and deposits which were close to each other were checked manually by browsing all data to ensure if the deposit was duplicate or not. Total duplicates found were 38. At this point, duplicates were marked and left to the database.
FODD dataset Structure of the FODD dataset was simple; only one table with comma-separated values in the fields. This database was processed and inserted to the new database structure. Some adjustments had to be made to fit data to the vocabularies. There were 345 deposits in this dataset.
Comparison between FODD and Migration database Before the cleaning of the Migration database was made, it was checked which deposits were also included in the FODD dataset. 110 deposits were connected by same coordinates and 131 by name/alternative name. Rest of the deposits were connected spatially with 500m buffer. Some of the FODD deposits were not in the Migration database, these would be added later.
Processing Migration database All comma-separated values were separated and processed into the new data model structure. There were numerous fields with these kinds of values, so this process will not be described in detail. Many of the fields contained descriptive text. Text was copied to the Comments -fields in the new database. Values were picked from the text and added to the fields with vocabularies. Those tables and fields which did not have place in the new database structure were removed (for example Whole rock geochemistry, Petrophysics, Ore composition, Zoning, Ore mineral composition, Stable isotopes, Fluid etc.) It has to be noted that before removing these fields it was checked that they didn’t contain any other information e.g. minerals, intersections. If so, this data was copied to the right place in database. There were also fields which contained information available from other sources: municipality, notes about location, geological units (group/formation), protection areas, etc. These were also removed. Following are some notes about migration; they do not contain every process step and description of the field sources, but rather show the problems with migration with few examples and some rules used for simplification and fitting the data. Because the most of the fields had to be processed manually, case-by-case, it was impossible to write down every process step and decisions made for simplification, combining or removing data.
EarthResource Orientation: New data model doesn’t contain field for strike, only for dip azimuth and dip. If there was information only about strike, this was put to Dimension_Comments -field. Many of the deposits contained orientation information as text, not as numbers. Conversion to the Migration database was done in the following way: • Almost vertical/subvertical=85 • Almost horizontal/subhorizontal=5 • Conversion from N, S, W, E, NE, NNE etc. to numbers. • If dip azimuth or dip was reported as range, average value was calculated, e.g. 10-30 ->20 OccurrenceType values were given to deposits using these rules: • If there has been mining activity, type is ‘Deposit’. Exception is when mine status is ‘Historic’ or ‘Closed’ and there has been mining activity less than 3 years (in these cases value is ‘Occurrence’). If there has been only test mining, value is ‘Occurrence’. • If resource estimate is made, value cannot be ’Prospect’. It is ‘Deposit’ if main commodity’s Importance value is 1-2 and ‘Occurrence’ if Importance value is 3-5. • If there is no resource estimate or mining activity, value is ‘Prospect’.
HolderHistory Current holders of the deposits were updated in ArcMap using land tenure polygons from Tukes.
Commodity Main commodity (Rank=1) was selected by the original database. Other commodities were picked from the other fields (e.g. Resources, Intersections).
MineralDepositModel There were comments about genetic type in four different fields in the source data. General description of geological setting were removed, genetic types were gathered to one table. Nomenclature used was very diverse; it was quite hard to fit deposit types to the vocabulary. Some challenges caused also the fact that the vocabulary changed couple of times during migration so simplification and fitting the data was done few times. Because of that it is very likely that there are some errors and wrong types in these fields. These will be corrected during update process.
ExplorationActivity and Intersections Exploration activity data was in Exploration –table. All activities in Au-Zn database were listed in one field so these had to be separated and all reference numbers had to be picked from the text. Drilling information was in separate column, but intersections were mostly in Resources-table. These were connected to exploration records. If the deposit had been drilled several times, it was hard to connect intersections to the right drilling record since this information was not usually available.
RockMaterial In Ni+PGE+Cu+U+IndMin –database host rock data was originally as separate records. Au-Zn -database was more difficult to process. Major and minor host rocks had to be processed manually, since majority of the information was not comma-separated. Reference numbers were in brackets, in most of the case at the end of the record, but in some cases in the middle of the text. It was important to keep these reference numbers connected to the deposit and rock types. Every one of rock types were separated to own row containing also fields for Deposit_ID, Proportion, References and Comments. Since comments were originally in one row, same comment was repeated for every rock type. Some of these comments were cleaned so that comment was pointing only to rock type in question. But majority of comments were left untouched (in the final stage, there was 2508 rock types, so it was impossible to go through every record). Same problem with comments was repeated with other tables too. New database structure is build so that RockMaterial is the main table in Geology-section. This caused some problems because source data was build differently; there were list of minerals, alteration, metamorphism etc. but records were not connected to any particular rock type. Only exception was Industrial Mineral database where minerals were connected to host rocks. It would have produced false data if all minerals, alteration etc. had been connected to all rock types in the deposit. The connection was made using following rules. Exception were those records where it was clearly stated in the comment field that this certain type of alteration (for example) occur only in certain kind of rock type Mineral: All ore and gangue minerals were connected to only one rock type which was presumably the main host rock type. In some cases host rock was not mentioned at all; in these cases host rock got value “NA”. Alteration: Alteration data was connected to only one rock type which was presumably the main host rock type. MetamorphicDescription: Metamorphic data was connected to all rock types of the deposit. RockStructure: Structure data was connected to only one rock type which was presumably the main host rock type. RockTexture: Texture data was connected to only one rock type which was presumably the main host rock type. GeologicalAge: Age data was connected to all rock types of the deposit. Only veins were excluded. RadiometricAge: Radiometric age data was connected to only one rock type.
Mineral Three different mineral-tables were combined (Major_ore_mineral, Minor_ore_mineral and Other_minerals). MineralProportion and OreMineral -fields were updated according to the source table. Mineral names were cleaned to fit MineralName-vocabulary. Au-Zn -database had different structure: they contained comma-separated text fields Major_opaques, Minor_opaques and Gangue. Siting of gold- and Fineness -fields were combined to Mineral table. Also fields Ore types, Primary textures, Secondary textures, Ore fabric were added to Mineral-table and MineralAppearance-table.
OreMeasure (Resources) Resource and reserve data was in combined Resources_Mining table along with mining activity data. This table contained also the best intersections which were separated to Intersections table. Following example of one record shows the starting point with the resource data: Down to 1000 m, measured+indicated+inferred resources (cut-off 0.2 % Ni ) [3]: 432 Mt @ 0.29 % Ni, 0.45 % Cu and 0.39 ppm PGE+Au. Cut-off 0.15 % Ni [4]: 89 Mt @ 0.26 % Ni, 0.4 % Cu and 0.6 ppm PGE+Au All of these records had first to be processed to the format x Mt @ x % Commodity1, x ppm Commodity2.... Original field contained all sorts of information about resource categories, cut-offs, name of the ore body, references etc. This information was copied to right fields and every resource estimate was separated to its own row. After that tonnes and grades were separated to own fields and data was imported to the new structure. Majority of the work had to be done manually.
Production Since databases didn’t contain detailed production data, this data was imported from excel-sheets compiled by K. Puustinen (http://weppi.gtk.fi/aineistot/kaivosteollisuus/). Mines were compared by name and spatially to connect production data to the deposits. Some of the historical mines were not in the database; these were left out from the database as also the historical mines which currently are located in Russia. Production excel sheets were processed and inserted to database. Total production numbers were calculated from the source data and these were compared to the total production numbers which were already in the Migration database. If there were differences in the values, data was checked and corrected.
References References were collected from original databases where they were marked as numbers in brackets within the text. Each database contained Reference_IDs as running number per deposit. With the original numbers, records in the Migration database were connected to the right source references. Links to source references were repaired; arkisto.gtk.fi addresses were changed to tupa.gtk.fi.
Figures Original databases contained links to the figures. Some of these links did not work; all links were checked and corrected. All figures were collected to one folder (in Master_Geodata), which contains subfolders for each deposit. Links to figures were re-built to point these subfolders.
Combining FODD and Migration database After data in the Migration database was processed, all resource and production numbers were once more compared with FODD data. Some deposit that were in FODD, were not in the Migration database. These 112 deposits were added. Deposit_IDs to the FODD deposits were created (8000+Deposit_ID). Combination of these two databases was easy, because FODD database was already converted to the new data model.
Combining old Ore database The last step was to compare the Migration database with the old Ore deposit database. Comparison was made by names and spatially. 235 prospects which were not in the Migration database were found in the old Ore deposit database. These were added to the database and Deposit_IDs were created (9000+Deposit_ID). Since the database structure was properly done in the old Ore database, combination of the data was quite easy.
Finalization After all deposits were in the database, the whole dataset was checked. Resource and production numbers were checked. Duplicate deposits were combined together and duplicate records in all other tables were removed. After all duplicates were removed; primary keys used during migration (database dependent PK) were replaced with running numbers. Last step was to fit data to vocabularies. In future all deposit data will be checked and updated by updating teams. |