A Durov diagram projects the cation and anion triangles onto a central square, classifying each sample by its dominant cation and dominant anion — and the layout makes evolution trends easy to see.
How it's built
Two ternary triangles feed a central square: the cation triangle sets the sample's horizontal position, the anion triangle its vertical position. The square is gridded into cells, so a point lands in a cell defined by its dominant cation and dominant anion. The expanded version adds two panels — pH below the square and TDS to its right — each sharing one of the square’s axes.
The axes, one by one
The Durov packs six compositional axes, two projected Cartesian axes and (in the expanded version) two more parameter axes into one figure. Read in order, they are:
The cation ternary — three axes. Each side of the top triangle is one cation group as a percentage of total cations in milliequivalents: calcium (Ca), magnesium (Mg) and sodium-plus-potassium (Na+K, almost always combined). The three are normalised to 100 %, so any two fix the third — a sample with 60 % Ca and 30 % Mg is necessarily 10 % Na+K. Which ion sits at which vertex is not universal; some software reorders them, so read the labels rather than assuming.
The anion ternary — three axes. The same idea for the anions along the left triangle: chloride (Cl), sulfate (SO4) and bicarbonate (HCO3, usually with carbonate CO3 folded in), each as a percentage of total anions in meq, again summing to 100 %.
The central square — two projected axes. This is where the reading happens, and it is a projection, not an independent plot. The cation triangle projects straight down, so cation composition sets the sample's horizontal position in the square; the anion triangle projects across, so anion composition sets its vertical position. Where the two projections cross is the water type, and samples that cluster in the square share a chemical character. One consequence worth knowing: because each point is projected perpendicular to its triangle's base, the third (apex) ion's information is flattened out in the square — the square faithfully carries two of the three ions from each triangle, not all three. That is exactly why the full triangles are kept alongside it, and why some workers deliberately reorder the vertices to put the ion they care about onto a base axis.
The pH panel — one axis (expanded Durov). The expanded version hangs a panel below the square that shares the square's horizontal, cation-driven position and plots pH down its vertical axis. So a point's left–right position still means cations, and how far it drops into this panel reads acidity.
The TDS panel — one axis (expanded Durov). Symmetrically, a panel to the right of the square shares the square's vertical, anion-driven position and plots total dissolved solids along its horizontal axis (some tools substitute electrical conductivity). Left–right in this panel means salinity. Together the two panels are what let a Durov carry composition, acidity and concentration in a single figure — its main advantage over a Piper, at the cost of taking a moment longer to read.
What to read
- Position in the square = water type. The top-left corner tends toward Ca–HCO₃ (fresh recharge); the opposite corner tends toward Na–Cl (evolved or saline). Movement along that diagonal often traces groundwater evolution.
- The 3×3 grid = process hints. The cells map dominant-cation × dominant-anion combinations, so clusters and trends across the grid point to mixing, dissolution or ion exchange rather than just labelling one water.
- Spread across the square = variability. A tight cluster is a chemically uniform dataset; a spread that crosses several cells flags mixing of distinct water types.
Watch for
Durov packs both ternaries into one figure, so it rewards a clean, well-grouped dataset — colour by group before reading it. And like any classification plot it describes proportions, not concentration, so check salinity separately.