Quick Q&A: What GRACE can and cannot Do

Context: The GRACE and GRACE-FO (follow-on) satellites are like a pair of scales in orbit: by tracking micrometre changes in the push and pull between two satellites, they map month‑to‑month changes in Earth’s gravity field, which are converted to surface mass change and expressed as ‘equivalent water height’ over regions a few hundred kilometres across. They have measured losses from massive ice sheets, shown where dry soils contribute to drought and changes in global groundwater resources.

What does GRACE actually measure?

• Measures: Tiny changes in gravity caused by shifts in mass, mostly from water moving between stores like groundwater, soil moisture, surface water (lakes and rivers), snow, and ice.
• Reports: These as “water-equivalent thickness” anomalies—basically, how much water height has changed compared to an average baseline period. It’s like weighing the extra or missing water on a huge scale.

Can GRACE measure groundwater directly?

• No. It captures the total water picture, so groundwater is estimated by subtracting modeled estimates of other components (like soil moisture or surface water) using tools like global land models. This subtraction introduces extra uncertainty from those models. Always cross-check with on-the-ground data, such as well measurements, for reliability.

What is the spatial and temporal resolution?

• Spatial: Covers huge areas, with an effective “footprint” of about 70,000–200,000 km², depending on the source and processing method (that’s an area bigger than Austria). Signals can “leak” or blur across edges, like near coasts or basin boundaries, making precise outlines tricky.
• Temporal: Provides monthly snapshots; it’s best for spotting long-term trends over years, not day-to-day changes.

Can GRACE detect local pumping or a new project?

• Usually not. Effects from a single bore or small development are too tiny and localised to show up clearly in GRACE’s broad view—they get lost in the noise. Even for larger basins, you need a strong, ongoing signal that’s bigger than the measurement uncertainty and lasts for years to detect it reliably.

Can GRACE tell me whether drying is caused by extraction?

• No, not on its own. It shows that water storage has changed, but not why. To figure out causes—like if human pumping is to blame—you need extra modeling that accounts for natural factors (e.g., rainfall patterns, climate cycles like El Niño), time delays in effects, and real data on water use. Even with downscaling GRACE can’t separate human from natural drivers in places like Australia. Scientists using GRACE data during the Millennium Drought in the Murray-Darling Basin just measured the overall drought levels across the whole Basin and didn’t try and report irrigation impacts.

What are common pitfalls in policy communications?

• Confusing “allocated” water (what’s permitted) with actual use (what’s really pumped).
• Treating “deficits” relative to a baseline as automatic proof of a specific cause, without deeper analysis.
• Mixing scales: GRACE sees the big picture (whole basins), while water licenses or projects are often local and small.
• Forgetting to account for uncertainties that build up from models, data gaps, or processing steps.
• Assuming timing alone (e.g., drying starts when pumping begins) proves causation, ignoring other factors like weather shifts.

When is GRACE most useful?

• Monitoring widespread droughts and how regions recover over time.
• Estimating long-term water balance trends in large basins.
• Tracking mass changes in ice sheets, glaciers, or sea levels.
• Verifying if computer models accurately capture big regional water swings.

When should GRACE not be used?

• For decisions about local water licenses or auditing individual projects—it’s too coarse for that.
• Short-term checks (like over a few months), where random noise can overpower real signals.
• Cases where the expected change (e.g., from small-scale pumping) is tinier than the method’s built-in uncertainty.

What does “below detection” mean here?

• It means if the water use (or its changes) is about 10 times smaller than GRACE’s trend uncertainty, you can’t reliably spot it at the basin level—it’s like trying to hear a whisper in a storm. A real local effect might still be there, but GRACE isn’t the tool to confirm or rule it out.

What is good practice for using GRACE in water management?

• Stick to its strengths: Apply it to large basins where the scale matches.
• Pair it with other data, like metered water use, well levels, and climate info (e.g., rainfall/evaporation anomalies, El Niño or Indian Ocean patterns).
• Use smart stats: Adjust for trends that aren’t steady, handle errors robustly, and test for time lags in effects.
• Always track and report uncertainties from every step, like modeling or filling data gaps.
• Be upfront about limitations: Say something like, “GRACE is great for tracking broad trends, but it doesn’t pinpoint causes at a local level or for small changes.”