Groundwater Science, Geochemistry and Data Analysis

Science beneath the surface

Water moves, even deep beneath the surface.

We share practical expertise in groundwater and geochemical modeling (MODFLOW, PHREEQC) for environmental and water resource assessment, along with data analysis using Python, AI, and Vibe Coding.
Exploring the world of water through science, code, and diverse perspectives—this is a platform to decipher natural phenomena from data.

SCIENCE · CODE · INSIGHT · WATER

Groundwater Geochemistry Hydrology AI & Python MODFLOW PHREEQC Geography Geothermal Science

About This Site

DeepFlows is a site run by a researcher specializing in groundwater science and geochemistry, dedicated to explaining the science and documenting the practice of data analysis. Rainwater seeps underground, flows slowly through aquifers, and eventually mixes with seawater along the coast. We decipher these "invisible movements of water," one step at a time, using numerical simulation and code.

This is not about memorizing textbook facts. From groundwater flow analysis with MODFLOW, to modeling water–rock chemical reactions with PHREEQC, to data visualization with Python, we explain the very methods used in real research—including the dead ends and the trial-and-error that rarely make it into published papers. The content is written for students learning hydrogeology, engineers in the water-resource and environmental fields, and anyone curious about numerical analysis and data science.

The figures and code on this site are built with an AI assistant (Claude Code). Scripts that compute molecular geometry and draw the figures, cross-checks of every number against its primary source, read-throughs of finished drafts—much of the hands-on work is delegated. What is not: deciding what to write, how to explain it, and whether a number can be trusted. Those remain entirely the author's. The rule that nothing goes in unless it can be traced to a source has not changed either. How to work alongside AI is itself one of this site's subjects.

📘

Introduction to Groundwater Science

From core concepts—aquifers, hydraulic conductivity, recharge—to time-series analysis of freshwater lenses, tidal response, and how earthquakes and volcanoes move groundwater. 18 parts so far, in Japanese and English.

⚗️

PHREEQC Geochemical Modeling

From the basics of water-quality calculation to reaction-path modeling and inverse modeling. A 20-part series explaining geochemical simulation with real examples and code.

🗒️

Field Notes

Why the hot springs of Beppu carry the chemistry they do. What the terrain at Sekigahara did to history. Taming partial differential equations; writing code alongside an AI. Essays that don't fit the tutorial format.

⚛️

The Elements

Why bed hair straightens with water. Why the Japanese word for lightning literally means "the rice plant's spouse." Science a high-school reader can follow, with figures that move—hydrogen, carbon, oxygen, nitrogen so far. Written in Japanese.＞ Start from the prologue

🧮

Online Calculators

Darcy's law, pumping tests, evapotranspiration, charge balance and water-chemistry unit conversion, worked out in your browser with nothing to install. Each tool cites its equation and states where that equation stops being valid.＞ Open Calc

Author: Heejun Yang, Ph.D.

A researcher specializing in hydrology, groundwater flow analysis, and geochemistry. Sharing the methods built up through years of research, in the plainest language possible. ＞ See full profile and publications

☀️ Featured Series

PHREEQC Tutorial

A complete guide to geochemical modeling, deciphering chemical reactions between water and rock. From installation to reaction path modeling.

READ IN ENGLISH > 日本語で読む >

Groundwater Science

Practical know-how for groundwater flow analysis and hydrothermal geochemical simulations using MODFLOW and FloPy.

READ IN ENGLISH > 日本語で読む >

Volcanoes and Groundwater — How Magma Type Decides Water Chemistry and Hot Springs (Basalt, Andesite, Rhyolite with PHREEQC) | Groundwater Science #18

Kusatsu discharges water at pH 2. Beppu discharges neutral chloride water. The deep waters of Shiraoi are soft and bicarbonate-rich. Why are the waters of volcanic districts so different from one another? The answer is the silica content of the magma. It fixes the viscosity, the eruption style and the shape of the aquifer; it fixes the rock-forming minerals and what weathering releases; it fixes the volatiles and the type of hot spring. This article dissolves basalt, andesite and rhyolite in PHREEQC, plots Japanese hot springs on a Cl–SO₄–HCO₃ diagram, and ties together what #2, #3, #11, #12 and #16 built.

Aug 11, 2026

DeepFlows

Earthquakes and Groundwater — How a Well Becomes the Earth’s Strain Meter (Poroelasticity) | Groundwater Science #17

At the instant of an earthquake the water level in a well steps up or down. Months before an eruption it drifts downward while oscillating. Weeks after a large earthquake, spring and stream discharge rises by an order of magnitude. This article works out how a groundwater level records crustal volumetric strain (linear poroelasticity, Δh = −Wε·Δε) through three studies: 19 wells in Hokkaido, the 2000 eruption of Usu volcano, and the 1989 Loma Prieta earthquake. It extends the view built in #6 (barometric response) and #9 (tidal response) to the strain of the crust itself.

Aug 5, 2026

DeepFlows

Geothermometry: Reading Reservoir Temperature from Hot-Spring Chemistry (Giggenbach Diagram & Silica Geothermometer) | Groundwater Science #16

Estimate the temperature of a geothermal reservoir kilometers underground — without drilling — from the chemistry of hot-spring water alone. Starting from a single principle (water–mineral equilibrium is temperature-dependent), we build up the Giggenbach Na-K-Mg diagram and the silica geothermometer, then test ‘how reliable is it, really’ at three hot springs: Hatchobaru (agreement), Shiraoi (mixing), and Bugok (disagreement). We reproduce the Shiraoi quartz temperature of 151°C in PHREEQC and cross-check against the public data of Shigeno (2011). The temperature volume of the ‘water tells its own history’ trilogy that follows #14 (isotopes) and #15 (maturation).

Aug 1, 2026

DeepFlows

Silicate Weathering and Groundwater Residence Time — Modeling How Spring-Water Chemistry “Matures” with PHREEQC | Groundwater Science #15

How many years does it take for groundwater to mature chemically? Using the Kirishima volcanic springs (residence times 1–58 years), we reproduce primary-mineral dissolution and secondary-mineral precipitation (gibbsite → kaolinite → Ca-smectite) in PHREEQC and cross-check against the public data of Ide et al. (2018). We read the “maturation clock” by which dissolved silica saturates and switches phases at about 20 years — and why fully quantifying it requires reactive transport.

Jul 30, 2026

DeepFlows

How Old Is Groundwater? — Reading Recharge and Residence Time with Isotopes, Tritium and ⁸⁵Kr | Groundwater Science #14

Groundwater records both where it was born (recharge source) and when (age). We review stable isotopes δ¹⁸O and δD, the meteoric water line and the altitude effect, and age dating with tritium, CFCs, SF₆ and ⁸⁵Kr — then read the stagnant, old water of Kumamoto (Kagabu 2017) and Kirishima (Ide).

Jul 29, 2026

DeepFlows

Why Arsenic Dissolves into Groundwater — Redox in Kumamoto’s Aquifer, Read with PHREEQC | Groundwater Science #13

In parts of Kumamoto, arsenic (As) exceeds the drinking-water limit naturally. The key is redox. Following Hossain et al. (2016), we reproduce it in PHREEQC — the redox sequence of electron acceptors, the pe/Eh control on arsenic speciation, and the release of arsenic by reductive dissolution of iron oxide.

Jul 28, 2026

DeepFlows

Why Fluoride Builds Up in Groundwater — Reproducing Kumamoto’s High-Fluoride Water with PHREEQC | Groundwater Science #12

In parts of Kumamoto, groundwater exceeds the drinking-water limit for fluoride (F⁻) naturally. Cation exchange turns Ca-HCO₃ water into high-pH Na-HCO₃ water, and that environment mobilizes fluoride from minerals. Following Hossain et al. (2016), we reproduce the process in PHREEQC — saturation indices, exchange, and the fluorite cap. This is inorganic fluoride, not PFAS.

Jul 27, 2026

DeepFlows

How Groundwater Chemistry Forms — Major Ions and PHREEQC | Groundwater Science #11

Why does groundwater differ in taste and composition from place to place? The origins of the major ions, the four mechanisms of water-quality formation (dissolution, cation exchange, evaporative concentration, mixing), and a PHREEQC calculation of rainwater dissolving calcite into a Ca-HCO₃ water — compared across river, groundwater, seawater and hot spring on a Piper diagram.

Jul 26, 2026

DeepFlows

What Groundwater Time-Series Analysis Reveals — FFT, Cross-Correlation and the Tidal Method | Groundwater Science #10

From a single water-level record, read the hidden properties of an aquifer. FFT, cross-correlation, cross-wavelet, the tidal method and lag-based regression — the tools built in #6–#9, organized into a practical guide of which method answers which question, revisited through Beppu, Minami-Daito and Tonle Sap.

Jul 25, 2026

DeepFlows

Introduction to PHREEQC #20: Reading Mineral “Winners and Losers” with the Equilibrium Constant (log K) and Gibbs Free Energy — Explaining Basalt-CO2 Secondary Minerals Thermodynamically

From the equilibrium constant (log K), Gibbs free energy (ΔG) and the saturation index (SI), read off which minerals precipitate in basalt-CO2 reactions. Add and subtract reaction log K values with Hess’s law, and use real PHREEQC/Thermoddem data (reproductions of Satake et al. 2025 and CarbFix) to explain calcite precipitation and smectite competition numerically.

Jul 16, 2026

DeepFlows