Seismic Performance of New Zealand Homes: Lessons from Recent Research
- Jan 13
- 7 min read
New Zealand’s light timber-framed (LTF) houses have long been designed to meet NZS 3604 standards. While these homes are generally life-safe, recent research and testing show that minimum code compliance may not be enough to prevent earthquake damage, particularly in modern homes with complex layouts, large glazing areas, or hybrid bracing systems.
Lessons from the past
Following the Canterbury earthquakes, NZS 3604 (2011) was updated. But the update was already underway before those events and didn’t fully capture the lessons from them. The revision made only incremental changes, like adjusting bracing demand and distribution rules, while still relying on the same P21 test and bracing unit approach. Research since then has shown that houses built to the minimum 3604 provisions can still exceed code drift limits, meaning the standard remains insufficient for controlling earthquake damage. Christchurch Earthquake (Canterbury: Magnitude 6.3) - Hundreds of buildings in the centre of the city have been demolished. The disaster is estimated to cost insurers $10-$20 billion.
At DTCE, we have been closely monitoring research into the seismic performance of New Zealand homes to understand how modern houses behave under earthquake loads.
The paper (“Seismic Performance and Loss Assessment of Light Timber Frame Residential Houses in New Zealand: State of the Art”) was written to provide a comprehensive review of the seismic performance of light timber-framed (LTF) houses. After the Canterbury earthquakes, there was strong interest in understanding how these houses behaved, where weaknesses occurred, and what improvements could be made.
General concerns with current design practice
Member-by-member approach:
NZS 3604 and BRANZ testing give reliable bracing capacities for individual walls and elements, but they don’t fully capture whole-house behaviour under seismic loading.
Observed issues in earthquakes: Even code-compliant houses showed uneven bracing distribution, torsional response, and foundation problems, which aren’t obvious when designing element by element
Load paths and interaction: Real houses act as a system. Diaphragms, connections, and foundations interact in ways not addressed by per-member design.
Why NZS 3604 Isn't Enough?
Before looking at specific test results and guidance, it’s important to understand how current standards perform in practice. While NZS 3604 provides a framework for designing light timber-framed houses, recent research and testing have highlighted areas where the standard may not fully protect homes against earthquake damage. The following section summarises the main issues and what the latest findings mean for modern residential design.
While NZS 1170.5 permits a storey drift of up to 2.5% at the ultimate limit state, test results show that plasterboard bracing walls perform effectively only up to about 1% drift, beyond which they degrade rapidly. As a result, when subjected to full design-level earthquake demand, the system is expected to displace to around 3% drift, exceeding both its effective capacity and the code drift limit.
So, BRANZ issued design guidance to help engineers go beyond the prescriptive rules.
Summary and Findings of this design guidance provided by BRANZ on the seismic design clauses of NZS 3604:
Due to the 3% drift, the minimum seismic bracing provision as per NZS 3604 would appear to need to increase by 50% so that the deflection requirement of the current seismic loading standard NZS 1170.5 is satisfied.
It is suggested that the effects of allowable irregular bracing arrangements within the scope of NZS 3604 on the seismic bracing requirement be studied. This is because the torsional effect is likely to lead to a further increase in seismic bracing demand.
Experimental Studies on Plasterboard Bracing Walls
TEST NO.1: Performance of walls with large window or door openings
To better understand how plasterboard walls perform under seismic loading, a series of experimental tests was conducted. These tests examined how walls with openings, such as windows and doors, respond to earthquake-type forces, and how construction details and connection methods affect overall performance.
Tested 10 long plasterboard walls with openings under reverse cyclic loads (P21 test style).
Looked at both windows and doors, and whether hold-downs were provided at the edges of openings.
Result:
1. Where plasterboard sheets join doesn’t matter much. Performance was very similar, whether:
Sheets were joined 300 mm or more away from an opening (Fig. 1a), or
Sheet edges were cut right at the edge of the opening (Fig. 1b).
So, in practice, both construction details gave comparable racking strength.
2. Two major deformation mechanisms:
Rocking of the entire panel (like the wall tipping back and forth on its base) → contributed 60–100% of total deformation.
Sheet rotation relative to the frame (from nail/fastener slip) → contributed 10–50%.
Findings:
Most wall drift comes not from plasterboard bending or crushing. This means that by improving fixings and connections, we can fully utilize the capacity of the wall. (See figure 2)
TEST NO.2: Single-storey House
Building on the wall-level studies, a full single-storey house was tested to observe system-level behaviour. This test examined how an entire structure responds to seismic forces, including how walls, ceilings, and connections interact under both dynamic and cyclic loading. Understanding the performance of the house as a whole helps identify potential weaknesses that aren’t visible when looking at individual elements alone.
Test house: a 1990s low-cost Fletcher home, plasterboard walls, fibre-cement cladding, no hold-downs.
Tests: free vibration + cyclic racking (jacks applied force to the ceiling).
Result:
Dynamic properties:
Fundamental natural frequency: 20.8 Hz → period = 0.05 s (a very stiff structure).
Equivalent damping: 8.2% critical damping (above the 5% often assumed in code).
Findings:
The house dissipates more energy per cycle than the code assumes.
In theory, this makes it less vulnerable to earthquake shaking, because it can “soak up” more energy before the damage escalates.
This means the house behaved as a system, not just a collection of isolated walls.
TEST NO.3: Repair of a single-room, one-storey structure about 2.4 m high
To explore repair strategies and their effectiveness, a single-room, one-storey structure was subjected to damage, repair, and retesting cycles. This approach allowed researchers to compare different repair methods: from simple cosmetic fixes to full strengthening, and understand how each method restores or improves bracing performance.
The key idea: damage → repair → retest → compare stiffness curves
In this model, walls and ceiling are lined with plasterboard. Most walls are standard plasterboard, no hold-downs. One wall (“BP10”) = bracing plasterboard with hold-downs acts like the “strong” wall in the setup.
The Test Setup (Figure 4)
Lateral loads are applied in cycles to simulate earthquake shaking.
Initial loading cycles: 1.65 mm, 3.92 mm, 7.29 mm displacements — enough to push it into early plastic deformation but not collapse
Phase | Repairs | Result |
|---|---|---|
| Cracks patched, plasterboard tidied up (think stopping/taping/painting). | Moderate improvement in stiffness, but not back to as-built. |
| Same cosmetic repair, plus extra drywall screws added between existing screws. | Surprisingly, little improvement beyond cosmetic-only repair. Shows that adding screws to already damaged boards does not add much. |
|
| This was the most effective repair, restoring and even strengthening the building’s bracing capacity. |
|
| Shows that the selected repair method significantly affects outcomes, with patching and strengthening performing very differently. |
DTCE's Perspective
At DTCE, we are seeing increasing complexity in residential builds:
Large glazing areas
Architectural forms with irregular layouts
Hybrid bracing solutions
To achieve truly resilient outcomes in modern residential builds, we consistently adopt a holistic, engineering approach:
Bracing Design: We go beyond the minimum requirements of NZS 3604, designing bracing systems that suit each home’s unique layout and materials.
Holistic Load Path Review: We assess the entire structural system to ensure loads are effectively distributed through walls, diaphragms, and foundations.
Connection and Diaphragm Strengthening: Hold-downs, anchors, and diaphragm enhancements are incorporated where needed to improve overall seismic performance.
Early Assessment of Torsional Behaviour: Irregular layouts can create torsional effects; we evaluate these early in the design process to mitigate drift and uneven wall response.
As seismic expectations evolve, so must our design approach.
In conclusion
New Zealand’s typical light timber-framed house has proven life-safe, but not necessarily damage-resistant. Research since Canterbury has demonstrated:
System-level behaviour matters
Drift performance, not just strength, is critical
Hold-downs and connection detailing can greatly improve outcomes
Minimum bracing may no longer be sufficient
For Homeowners, Designers, and Insurers, this research highlights the importance of engineering solutions, not just prescriptive compliance.
If you’re considering a new build, home extension, or seismic assessment in New Zealand, you might be wondering: How do seismic assessments work for homes? And how do I choose the right structural engineer?
Seismic assessments evaluate how your home would respond to earthquake forces, identifying areas of potential weakness and recommending improvements. Choosing the right structural engineer is key: look for someone who goes beyond minimum code compliance and focuses on real-world building performance.
If you have questions about your home or want a fresh perspective on its seismic performance, feel free to reach out, we’re here to help.
Sources:
Comments