Operational stormwater performance of building integrated green walls

Technical Summary: Operational Stormwater Performance of Building-Integrated Green Walls

Problem Statement
Urban densification and intensifying convective rainfall are increasing pressure on drainage systems. While Nature-based Solutions (NBS) like green walls offer a decentralized approach to managing roof runoff by utilizing facade space, quantitative design guidance remains immature. Existing literature reports single-event retention rates ranging widely from 55% to 100%, but these values are difficult to transfer to design contexts because they often lack consistent reporting of initial substrate moisture, storm magnitude, and antecedent dry periods. Furthermore, many studies rely on "dry-start" assumptions (fully empty substrate) which may represent favorable, non-operational conditions rather than realistic scenarios where storms follow short dry spells. There is a critical gap in understanding how antecedent storage states translate into available retention capacity across different wall architectures and consecutive storm events.

Methodology
This study employs a laboratory-to-model-to-design workflow to evaluate roof-runoff-fed green walls.

• Experimental Setup: Full-scale experiments were conducted at the Green Sponge Buildings (GSB) Laboratory in Norway on two contrasting modular green wall systems:
  • GW1 (Tank-type): A commercial system with built-in reservoirs and mineral substrate, exhibiting "fill-and-spill" behavior.
  • GW2 (Porous-substrate): A system with absorbent precast elements forming a continuous, free-draining surface, allowing throughflow before full saturation.
  • Both walls were instrumented with flow meters and load cells (weighing the walls at 1 Hz) to independently verify retained volumes and evapotranspiration, addressing measurement uncertainties common in outflow-only studies.
• Modeling: A parsimonious two-zone event model (retention and detention zones) was developed at a one-minute timestep. The model incorporates a wall-to-roof ratio ($\lambda = A_{wall}/A_{roof}$) as the primary design variable.
• Simulation Scenarios:
  • Model Evaluation: The model was calibrated against observed GSB Lab events for both wall types.
  • Design Storms: Single-event simulations used Oslo IDF (Intensity-Duration-Frequency) data for return periods of 2 to 100 years.
  • Repeated Storms: Chains of five identical 2-year storms were simulated with inter-event gaps ranging from 3 hours to 15 days to assess retention recovery via evapotranspiration (using GLEAM data and FAO-56 crop coefficients).
  • Uncertainty Analysis: A Latin Hypercube Sampling (LHS) approach ($N=500$) was used to quantify prediction uncertainty, sampling storm depth, antecedent dry periods, crop coefficients, and storage capacity.

Key Results

• Model Performance: The two-zone model accurately reproduced observed retained volumes for both wall types (GW1: NSE = 0.99; GW2: NSE = 0.97), validating the approach for distinguishing architecture-specific hydraulic behaviors.
• Architecture Differences:
  • GW1 (Tank): Exhibits high retention capacity ($h_{ret} = 38.0$ mm) with a "fill-and-spill" mechanism. It achieves near-full retention for frequent storms at modest wall-to-roof ratios.
  • GW2 (Porous): Has lower capacity ($h_{ret} = 11.25$ mm) and early drainage ($\theta = 0.30$). It provides partial retention but requires significantly larger wall areas to achieve high retention targets.
• Impact of Antecedent Conditions: The study found that the pre-storm storage fraction ($f_0$) is the dominant explanatory variable for retention variance ($R^2 \approx 0.97$).
  • Under realistic Oslo antecedent conditions (mean initial storage fractions of 0.87 for GW1 and 0.64 for GW2), dry-start assumptions significantly overstate performance.
  • For a 10-year storm at $\lambda = 1.0$, dry-start assumptions overstated median retention by 89 percentage points for the tank wall and 30 percentage points for the porous wall.
• Repeated Storms: Retention degrades rapidly in consecutive events with short inter-event gaps (e.g., 3–6 hours) because evapotranspiration cannot restore capacity between storms. GW2 degrades faster than GW1 due to its smaller storage volume.
• Design Implications:
  • To meet Oslo's requirement of retaining the first 10 mm of rainfall: GW1 requires a wall-to-roof ratio ($\lambda$) of $\approx 0.26$ (dry-start) or $\approx 0.54$ (50% pre-filled). GW2 requires $\lambda \approx 2.33$ (dry-start) and cannot reach near-full retention within practical facade ranges if pre-filled.
  • Peak flow reduction does not always mirror volume retention; once capacity is exceeded, the detention zone can still dampen peak flows even if volume retention is limited.

Significance and Claims
The paper argues that operational stormwater performance of green walls cannot be defined by a single retention percentage or substrate type alone. The authors claim that:

1. Antecedent State is Critical: Design and performance reporting must explicitly account for wall architecture, wall-to-roof ratio, and antecedent storage (pre-storm wetness). Relying solely on dry-start retention leads to significant optimism bias in design.
2. Architecture Dictates Function: Tank-type walls are suitable for near-full retention of frequent events, whereas porous walls with early drainage act primarily as partial-retention measures unless paired with additional storage (e.g., cisterns).
3. Design Guidance: The study provides a transferable workflow for sizing roof-runoff-fed green walls. It establishes that for frequent to moderate events, green walls can serve as effective building-integrated source controls, but they are not stand-alone protection against rare pluvial flood events due to the sharp increase in required facade area with storm depth.
4. Regulatory Reporting: The authors recommend that regulatory claims and performance reports use antecedent-conditioned percentiles or explicitly stated pre-fill assumptions rather than dry-start values, which should be treated only as upper bounds.

The study concludes that while green walls offer a viable pathway for retrofitting building envelopes for stormwater management, their operational efficacy is governed by the interplay of storage capacity, facade-to-roof ratios, and the realistic recovery of storage between events.