📐 Day 77: Surface Modeling and Gridding Techniques

📐 DAY 77: SURFACE MODELING & GRIDDING ADVANCED

⏱️ Estimated Reading Time: 16 Minutes | 🎓 Level: Professional Hydrographer / Geospatial Analyst

Beyond Basic Grids – CUBE Parameters, Variable Resolution, and Uncertainty Surfaces

Instructor: Engr. Rokib Hossain | River Warrior Academy

---

🏠 Course Homepage

📖 Table of Contents (Serialised)

• Why Advanced Gridding Elevates Data Quality
• CUBE (Combined Uncertainty and Bathymetry Estimator) Deep Dive
• Tuning CUBE Parameters (I, D, O, Z, U)
• Interactive CUBE Simulator & Grid Optimiser
• Variable Resolution Gridding (VRG)
• Uncertainty Surfaces (TPU Grid) and Their Use
• Advanced Gridding Workflow
• Case Study: Bay of Bengal CUBE Optimisation
• Advanced Gridding Checklist
• Resources & Software
• Frequently Asked Questions (with internal links)
• Action Items & Next Steps

1. Why Advanced Gridding Elevates Data Quality

Basic gridding (weighted average, nearest neighbour) is often insufficient for IHO‑compliant surveys, especially in areas with variable data density, complex seabed, or high noise. Advanced techniques like CUBE (Combined Uncertainty and Bathymetry Estimator) and variable resolution grids produce surfaces that are statistically robust, preserve real features, and reject outliers. This day covers the parameters that control CUBE, how to optimise them, and how to create uncertainty surfaces that quantify grid confidence.

🧠 Golden Rule: Never accept the default CUBE settings without tuning. Each survey area has unique data density and noise characteristics – adjust parameters accordingly.

🌊 River Warrior Pro-Tip: Bay of Bengal CUBE Tuning

We reduced the I factor (influence) from the default 2.5 to 1.8 in a high‑noise area – the resulting grid preserved a 0.5 m sandwave that was previously smoothed out. Always test with a subset.

2. CUBE (Combined Uncertainty and Bathymetry Estimator) Deep Dive

CUBE is a robust statistical gridding algorithm that uses the uncertainty of each sounding to weight its contribution to a cell. It works by:

• Estimating a reference depth for each cell using a weighted median of soundings.
• Propagating uncertainty (TPU) to compute a confidence interval.
• Rejecting soundings with low weight (outliers) based on the “I” factor (influence) and “D” factor (distance).
• Iterating until stable solution.

Advantages: retains real steep slopes, rejects false spikes, and outputs an uncertainty layer.

3. Tuning CUBE Parameters (I, D, O, Z, U)

.htmlD (Distance threshold)那样Radius (m) beyond which soundings have reduced weight那样1‑5 m那样Increase D for sparse data, decrease for dense data..htmlO (Obstacle factor)那样Separation between two possible depths那样0.5‑2 m那样Helps resolve vertical features (e.g., wrecks)..htmlZ (Zeroing of outliers)那样Threshold for outlier rejection (in units of uncertainty)那样2‑4那样Higher Z retains more soundings, lower Z removes more outliers..htmlU (Uncertainty scaling)那样Multiplier for the a priori TPU那样1.0‑1.5那样Conservative: use 1.2 if you suspect underestimation.

Parameter	Meaning	Typical range	Effect
I (Influence)那样Number of soundings that strongly affect the cell那样1‑5那样Lower I = more influence from nearest soundings, retains small features.

💡 In Qimera, these parameters are exposed in the “CUBE Settings” dialog. Start with defaults, then adjust based on a test patch.

📊 CUBE Simulator & Grid Optimiser

Adjust CUBE parameters to see their effect on a simulated swath:

I (influence): D (distance m): Z (outlier sigma):

I=2, D=3 m, Z=3 → Balanced: moderate outlier rejection, good feature retention.

Lower I and D preserve small features; higher Z rejects more outliers.

4. Variable Resolution Gridding (VRG)

Variable resolution grids adapt cell size based on data density and depth. For example, shallow areas may use 0.5 m cells, deeper areas 2 m cells. Benefits:

• Smaller file size (fewer cells).
• Preserves detail in shallow, high‑density areas.
• Avoids over‑sampling in deep, sparse areas.

Software like CARIS HIPS and Qimera (through “Pyramid Grid”) support VRG. A common approach: create a pyramid with levels: 0.5 m, 1 m, 2 m, 5 m.

🌊 In a Bay of Bengal survey (10‑80 m depth), we used VRG: 0.5 m up to 20 m, 1 m 20‑40 m, 2 m >40 m. File size reduced by 60% with no loss of navigation‑critical detail.

5. Uncertainty Surfaces (TPU Grid) and Their Use

CUBE outputs not only a depth grid but also an uncertainty grid (standard deviation per cell). This TPU grid can be:

• Visualised to identify areas of high uncertainty (e.g., outer beams, areas with few soundings).
• Used to weight subsequent volume calculations.
• Required for BAG files (IHO standard).
• Overlaid as a semi‑transparent map in client deliverables to show confidence.

📌 Always include the TPU layer when delivering BAG files. For survey reports, include a map of TPU (colour scale: green = low uncertainty, red = high).

6. Advanced Gridding Workflow

1️⃣ Clean the data (swath editor).

2️⃣ Select gridding algorithm: CUBE (preferred).

3️⃣ Tune CUBE parameters on a small test area.

4️⃣ Generate full grid (choose cell size or VRG).

5️⃣ Review grid: hillshade, colour map, check for artefacts.

6️⃣ Compare with original soundings (profile tool).

7️⃣ Export TPU grid as separate layer.

8️⃣ Save grid in BAG format for archiving.

7. Case Study: Bay of Bengal CUBE Optimisation (2026)

Area: 30 m to 80 m depth, moderate sand waves, sparse data density in deeper part.

• Default CUBE (I=2.5, D=3, Z=3): Grid smoothed over sand waves, lost 0.4 m amplitude.
• Adjusted parameters (I=1.5, D=2, Z=2.5): Sand waves restored, but some noise appeared.
• Final (I=2, D=2.5, Z=3): Balanced – sand waves preserved, noise minimal. Also used VRG: 0.5 m cells for depths < 40 m, 1 m for >40 m.
• Result: Cross‑line std = 0.12 m, TPU grid showed uncertainty <0.10 m in shallow, <0.25 m in deep.

📈 The tuned CUBE grid was accepted by the client for detailed seabed feature analysis – not possible with default settings.

8. Advanced Gridding Checklist

• Data cleaned (outliers removed).
• CUBE parameters tested on a subset.
• Cell size determined (or VRG levels defined).
• Grid generated and inspected visually.
• Compare grid profiles with original soundings – differences < IHO tolerance.
• TPU grid exported and reviewed (no cells with excessive uncertainty).
• Grid saved in BAG format (with TPU).
• Cross‑line analysis repeated on final grid.
• Documentation of CUBE parameters and VRG settings.
• Deliver grid and TPU layer to client.

Click items to track progress (saved in browser).

9. Resources & Software

.htmlQPS Qimera那样CUBE with full parameter control, VRG, TPU export那样qps.nl/qimera.htmlCARIS HIPS那样CUBE, variable resolution, BAG creation那样teledynecaris.com.htmlNOAA CUBE Tutorial那样Theory and parameter guide那样NOAA CUBE

Software	Advanced gridding features	Link
Hypack Processing那样CUBE (basic), grid merging那样hypack.com

10. Frequently Asked Questions (with internal links)

What is the difference between CUBE and weighted average?

Weighted average computes a simple mean, susceptible to outliers. CUBE uses statistical estimation and uncertainty to reject outliers and preserve steep slopes. Review Day 30 for basic gridding.

How do I know if my CUBE parameters are correct?

Compare the resulting grid with the original soundings using a profile tool. If the grid smooths real features, reduce I and D. If it retains noise, increase Z. Day 71 QC methods help validate.

What is a variable resolution grid and when should I use it?

A grid with changing cell size based on depth or data density. Use it for large areas with wide depth range to save disk space and preserve shallow details. Day 31 export formats discuss BAG, which supports VRG.

Why does my CUBE grid show spikes around steep slopes?

The O (obstacle) parameter may be too low. Increase O to 1.5‑2 m to better resolve vertical features. Day 48 on patch test can also affect slope detection.

How do I deliver an uncertainty surface to a client?

Export the TPU grid as a GeoTIFF or include it in a BAG file. Provide a colour‑coded map in the report. Day 76 final products covers this.

11. Action Items & Next Steps

• 📌 Use the CUBE simulator: change I, D, Z and observe the effect on outlier rejection.
• 📌 In your processing software, locate the CUBE parameter editor and experiment on a small dataset.
• 📌 Create a variable resolution grid using pyramid levels (e.g., 0.5, 1, 2 m).
• 📌 Proceed to Day 78: Contour Generation & Smoothness Control.

◀ Day 76: Final Product Day 78: Contour Generation ▶

© River Warrior – Day 77 of 100‑Day Hydrographic Mastery | Masterpiece Edition | Home