Uncovering a Major Flaw in the Flood Model

Evidence of a fatal flood model flaw that the Councils have kept covered up

Greater Wellington Regional Council (GWRC) and Upper Hutt City Council (UHCC) carried out a piece of modelling work which they called a “future case scenario”. They looked at the impact on flooding of 1,665 new houses on the Pinehaven hills, on land owned by Guildford Timber Company (GTC). Their modelling showed that the increase in flooding would be “minor ... less than 100mm increase in inundation depth across the catchment” (SKM:2010).

 

The community, concerned about the credibility of this finding, invited Graham Campbell, Manager of GWRC’s Flood Protection Department, to address a public meeting to explain how such a large development could have only a minor impact on flooding. The community also asked Mr Campbell to explain how this baseline flood model would work for assessing any future development on the Pinehaven hills by Guildford Timber Company for ‘hydraulic neutrality’.  Mr Campbell declined to answer these questions.

 

That public meeting, organised by the community, was held in Pinehaven in 2014. It marked the beginning of a 7-year saga in which the Councils (GWRC and UHCC) have continued to avoid addressing a “major flaw” in their Pinehaven flood model. We now know what that flaw is. Council’s base model misrepresents the existing catchment as highly impervious. Adding 1,665 new houses to a highly impervious catchment makes little difference to the huge volumes of stormwater already in the model before the extra housing is added.  

 

The reality, proven by field tests, is that the Pinehaven catchment, about 80% of which is forest and bush, is highly permeable. The forest acts like a giant sponge, soaking up large volumes of stormwater. Replacing the forest with urban development (large areas of impervious roads, roofs and driveways) would result in a significant increase in stormwater runoff and flooding. The Auckland Council stormwater design guide GD04:2015 comments that urban development on greenfield sites typically results in increases in stormwater runoff of 200% (for low density housing) to 500% (for medium density housing).

 

The purpose of ‘hydraulic neutrality’ rules is to ensure that developers manage the extra stormwater runoff created by their developments so that it doesn’t make flooding down-stream any worse than it was before the development. But for hydraulic neutrality rules to work, an accurate base model is needed of the existing situation before the development. Council’s base model of the Pinehaven catchment is grossly inflated because it incorrectly assumes the catchment to be highly impervious and therefore it is not a reliable model for assessing future developments for hydraulic neutrality. If Council’s Pinehaven flood model is used as the baseline model then hydraulic neutrality simply won’t happen.

 

For the past 7 years the Councils have avoided addressing this ‘major flaw’ in the Pinehaven flood model that is revealed in the ‘future case scenario’ modelling. Graham Campbell avoided talking about it (2014); the Beca auditor discovered it but failed to disclose and address it (2015); GWRC’s Pinehaven Floodplain Management Plan Hearing Panel ignored it (2016); the Focus Groups facilitated by UHCC allowed no discussion about it (2016); Commissioner McMahon completely glossed over it in the Plan Change 42 Hearing (2017); in the Environment Court, UHCC refused a mediation meeting about it with SOH, and the Envirnmenrt Court judges refused to allow an Alternative Dispute Resolution (ADR) process that would have enabled it to be properly addressed (2018/2019); the GWRC Whaitua Te Whanganui a Tara ignored it (2019); UHCC refused an offer of a workshop on it with SOH (2020); the three Commissioners at the Pinehaven Stream Improvements Hearing (2020) simply ruled it ‘out of scope’; and there was zero engagement from GWRC’s Hutt Valley Flood Management Subcomittee when SOH presented key evidence to them on it (2021).

 

The processes for developing and approving the Pinehaven Stream flood model and flood maps have been deeply flawed. The “major flaw” in the Pinehaven Stream flood model still exists. The system has failed to provide the much-touted hydraulic neutrality required for flood protection from any future Guildford Timber Company development on the Pinehaven hills for the future safety and wellbeing of the Pinehaven community.

Timeline - Development of the Pinehaven Stream Flood Model and Flood Maps

The Beca audit of the Pinehaven flood maps rated the ‘minor’ impact result of the ‘future case scenario’ as a “major flaw” in the flood model.  The audit report notes that when 1,665 new dwellings are added into the modelling “the flood volume does not increase. This is unexpected, as increasing the impervious area ... to reflect the development would be expected to reduce rainfall losses and increase runoff volume.” (Beca:2015, p26)

 

SOH has a copy of email correspondence between the Beca auditor, MWH and GWRC (obtained under the Official Information Act) which shows that the auditor discovered why there was no increase in flooding. In an email chain under the Subject heading “Pinehaven Stream hydrology - Existing and Future Development”, the auditor provided a detailed explanation of the apparent flaw in the modelling to MWH on 27 May 2015.  The auditor followed up with an email to MWH on Thursday 11 June 2015:

 

Beca:     

… please let me know whether … you … have found an answer to my query about the future development hydrology [why there isn’t the expected increase in flood volume in the post-development model compared with the pre-development model].

 

MWH:  

... the initial Loss and continuing losses used are the same in both [pre-development and post-development] models … [5mm Initial Rainfall Loss and 2mm/hr Continuing Rainfall Losses] … The Hydrol model doesn’t take any further excess out apart from the initial and continuing losses so that could explain why the volume is the same.

 

Beca:     

If the initial and continuing losses are the same in both models, then the flood volumes will be the same …

 

‘Rainfall loss’ is rain that is lost to such things as interception by trees and other vegetation, evaporation, and infiltration into the ground. Rainfall that isn’t lost becomes surface runoff. The gist of the above email chain conversation is that both the pre-development model (existing forest without the 1,665 new dwellings) and the post-development model (in which forest is replaced with roads, driveways and 1,665 new dwellings) have the same rainfall loss of 2mm/hr.

 

A rainfall loss of only 2mm/hr might be typical for a supermarket car park where a small amount of rainfall might be retained in shallow puddles on the ashphalt and evaporate away while most of the rainfall runs off. This might be appropriate in the post-development model but is certainly not a true representation of the pre-development situation of rainfall in the existing forested Pinehaven hills. Field tests in forest and bush areas on the Pinehaven hills consistently showed infiltration rates of 500mm to 900mm per hour (Alex Ross:2019).

 

However, the Beca auditor did not disclose this gross error in the Councils’ baseline flood model but erroneously reported the flood model and flood maps to be “fit for purpose”. This is a gross injustice, especially since the spurious ‘future case scenario’ was the main reason why the community requested the flood mapping audit in the first place.

 

Upper Hutt City Council (UHCC) says “... the Pinehaven flood modelling was tested through a robust process as part of Plan Change 42 to the District Plan. This process included the hearing ...  chaired by an independent Hearing Commissioner.” Greater Wellington Regional Council (GWRC) says the Pinehaven “flood maps have been peer reviewed and accepted by the environment court as ‘fit for purpose’”.  But there exists no declaration from the Environment Court that the Pinehaven flood maps are “fit for pupose”.  UHCC’s Plan Change 42 hearing process was not robust; the ‘independent’ Hearing Commissioner glossed over the evidence of flood map flaws, as the following brief summary points out.

 

The error in the ‘Future Case Scenario’ reveals that the baseline flood model is fatally flawed, but this error has been persistently avoided by the Councils and their consultants.

 

At the PC42 Hearing (2017) the Beca auditor claimed Jacobs’ reworking of the ‘Future Case Scenario’ (2016) fixed the error in the flood model. R J Hall’s peer-reviewed evidence (5 Nov 2019) shows Jacobs’ reworking of the model didn’t address the error, and the error in the baseline model still exists. R J Hall’s summary (2 Dec 2020) explains hydraulic neutrality will not happen for future development on the hills because of the error in the baseline model; Pinehaven residents and the environment will be exposed to greatly increased flood risk from future development on the Pinehaven hills due to the flawed baseline flood model.

 

It is not just properties in the flood zone that are at risk from Council’s flawed flood model. Because of the flawed model, the impact of future development on the hills will be completely unpredictable. There will be lots of dangerous, unmanaged stormwater runoff from future development on the hills that will affect properties in the foothills as well as in the flood zone.

 

Residents and properties in Silverstream will be affected too. The proposed streamworks will not protect residents in Silverstream from very large volumes of unmanaged runoff that the flawed flood model will allow from future development on the hills. Residents in foothill areas like Sylvan Way, Kiln Sreet, Amberly Gardens, Kurth Crescent, Field Sreet, Duncraig Street and Dunleith Grove will be just as vulnerable as residents in the foothills in Pinehaven.

 

In its recently adopted Long Term Plan 2021-2031, Upper Hutt City Council has allowed $15.2M for the Pinehaven Streamworks, but none of this is being spent on flood protection for Pinehaven! Pinehaven is getting zero flood protection! And in Silverstream, besides the Pinehaven Road and Sunbrae Drive road culvert replacements, there is no information in the Long Term Plan about what Council is doing for flood protection.

 

The ‘Future Case Scenario’ alone is enough to show that Council's baseline Pinehaven flood model and flood maps are fatally flawed, but there is other evidence as well: 

 

  • the gross mismatch of Council’s 100-year flood extents with eye-witness accounts of the December 1976 flood extents; 

  • R J Hall’s expert evidence for 100-year flood extents at 27 Elmslie Road, Pinehaven;

  • Save Our Hills (SOH) 4 case studies - avoided by UHCC, GWRC and their consultants at the audit ToR meeting (Feb 2015);

  • Save Our Hills (SOH) 4 case studies - ignored by GWRC’s Pinehaven FMP Hearing Panel (7 Apr 2016);

  • Focus Group meetings (Apr/May 2016) - the UHCC facilitator refused to allow any discussion of evidence of flaws in the Pinehaven flood model and flood maps;

  • peer reviewed expert evidence of multiple flaws in Council’s modelling of the flood event on 23 July 2009, the only event on which Council’s flood model is based; 

  • a gross mismatch of Council’s 1-in-25 year flood extents with actual flood extents in a 1-in-25 year flood that occurred in Pinehaven and Silverstream 8 December 2019.