Column ingestion FAQ

Common questions about preparing a column spreadsheet for Macrostrat. For the step-by-step guide, start with the quickstart guide; every sheet and field is described in the full specification.

Getting started

There are two. The simple template is a measured column with one position per unit and only the essential fields, and is the one the quickstart walks through; the expanded template has every field. Both are read the same way on upload. Their tabs follow the steps of the quickstart guide, their headers are coloured by whether a field is required, and each data tab is followed by an "(example)" tab showing what every field means, its format, and filled-in rows. A worked example (nine measured sections digitized from a published figure) shows a complete workbook.

A column is one outcrop or core site. A section is part of a column with one continuous age model; you only need more than one where the age model breaks, for example across a fault. A unit is one row of the units sheet: any interval of rock you can describe as a single chunk, from one bed to a whole formation.

Use measured for a measured section or core, where units are placed by measured height. Use composite for a composite or chronostratigraphic column, where units are placed in age order. Most field-measured data are measured columns. The older values section and column are still read the same way.

Yes. Give each a row on the columns sheet with its own col_id, and tag each unit with the col_id of the column it belongs to.

Units and positions

No. Every unit needs one position, the height of its base: position in the simple template, b_pos in the expanded one. Every unit's top is taken from the base of the unit above it. The topmost unit has nothing above it, so end the column with an empty closing row: a row with only col_id and a position giving the height of the top of the column. It is not a unit, but a b_int and b_prop on it date the top. In the expanded template, a t_pos on the topmost unit works instead, and is also how overlapping units are given. Row order doesn't matter, because units are sorted by height.

As small or large as makes sense for your data. A unit can be a single bed or a whole formation, as long as it can be described as one chunk. When in doubt, choose the simpler option and note any generalizations in comments.

A covered interval (rock that is there but not exposed) is a unit like any other. Name it covered, as its unit_name or in place of its lithology; or, if you can infer the lithology, such as shale for a recessive interval, write that and set covered to y (expanded template). A gap (missing rock) is not a unit: add a row named gap (as its unit_name) at the base of the missing interval, and no unit is created between it and the next row. Say what the gap is in comments. A gap does not need a new section_id unless the age model changes across it.

Each unit needs a strat_name (the formation or member), a unit_name (a local name such as "upper carbonate"), or both. If unit_name is blank, the strat_name is used as the unit's name; with neither, the unit is saved as "default". The template turns both cells red when a row has neither.

Lithology

Descriptive words first, then the rock name: cross-stratified sandstone. Separate several lithologies with ;, and put lesser ones in minor_lith (expanded template). For a range, list both end members: "light grey to black shale" becomes light grey, black shale. Attributes need a rock name to attach to; "flute casts" on its own is not a lithology.

Feel free to diverge a bit. Describe your rocks the way you would in a paper, using Macrostrat's lithology names where they fit. Words the ingester can't match are left out of the parsed lithology, with a warning, rather than stopping the upload. They are also how Macrostrat's lexicon gets better over time: they show us which terms people actually use and are missing. To see how your wording will be read before you submit, paste it into the lithology matcher. Interval names are the exception: they must match a Macrostrat interval exactly.

No. A unit that names a facies in its facies field takes the facies' lithologies when it has no lithology of its own, and the facies' environments when it has no environment. A unit's own cells always win, so fill them in only where a bed differs from its facies.

No. Set fill_values to y on the metadata sheet and blank cells take the value of the nearest unit below them (lithology, strat_name, unit_name, environment, facies and a few others). Enter a value at the base of the interval it applies to, and type none in a cell to stop it filling further up. In a core logged by depth, values fill downward instead, from the top of each interval.

Ages

Enter ages only at tie points: a unit's b_int (a geologic interval, such as Terreneuvian) and b_prop (where its base falls in that interval, from 0 at the oldest end to 1 at the youngest). Each section needs at least two tie points at different heights; ages in between are interpolated by height. Write your reasoning in an age_model_notes column.

It is read as 0, the oldest end of the interval. If you know the interval but not the position within it, an educated guess is better than a blank.

They are not extrapolated: units beyond the outermost tie points take the age of the nearest one. Add a tie point near the top and base of a section if their ages matter.

Give the correlated horizon, such as a datum bed, the same b_int and b_prop in every column. If a correlation falls inside a unit, split the unit at that height.

Not yet: ages are entered as intervals and proportions. Record dated horizons in comments or age_model_notes, and use them to choose your tie points.

Interval names must match Macrostrat's exactly. Look them up in Macrostrat's interval list.

Locations and sources

Estimate them from the paper's location map or satellite imagery, enter them as lat and lng in decimal degrees (west is negative), and note how they were obtained and how precise they are in the column's comments.

geom is for columns that cover an area (a POLYGON or MULTIPOLYGON) or follow a measured traverse (a LINESTRING or MULTILINESTRING), in well-known text. For a single location, use lat and lng instead; a POINT in geom is rejected.

Paste the original figure on the template's images tab, and name it there with its figure number and source (the tab is ignored on upload). Say in the metadata sheet's comments how heights were measured, for example "digitized from Fig. 11a with WebPlotDigitizer". This lets others see what was generalized.

Uploading

Upload it on the new column page ("Upload data") with dry run ticked. A dry run performs the whole ingest and then undoes it, so it reports errors without saving. Any signed-in user can run one; saving columns is limited to administrators.

Yes. A successful dry run opens the parsed column in the column editor, unsaved. If the workbook has several columns, each gets an "Open" button and an "Open in new tab" button, so you can look at them side by side. Edits made there stay in the page until you export them.