Green Infrastructure Capital and Operation and Maintenance Costs - City of Philadelphia Clean Waters Pilot Program Final Report

previous post summarized budget costs for Philadelphia's extensive green infrastructure program, showing budget costs of $568,00 per hectare, comparable to recent Ontario LID project tenders with an average cost of $575,000 per hectare.

The Philadelphia Water Department's has also reported extensively on green infrastructure costs and performance in their report Green City, Clean Waters Pilot Program Final Report. Highlights are presented below.

Green Infrastructure Capital Costs (Construction)

"The median construction cost per unit of impervious drainage area was $353,719/ac" - that equates to $872,000 per impervious hectare (2015 dollars).

"Median construction cost per unit of storage volume (Greened Acre) is $248,365/ac-in" - that equates to $2416 per cubic metre.

Overall costs appear to be increasing over time as shown in the following chart - to convert cost per acre to per hectare, multiply by 2.47 :

Green Infrastructure Construction Cost by Feature Type

Capital costs vary according to the type of green infrastructure (called GSI in Philadelphia). The following chart shows the variability in cost per managed impervious area for various types, suggesting some economies of scale for larger managed impervious areas.

The following chart shows the range of cost, median and average cost per managed impervious acre. A high variability in costs is shown from project to project.

Construction Cost by Loading Ratio / Efficiency

The cost efficiency of a green infrastructure project can vary according to its loading ratio, i.e., the relative size of the contributing runoff area to the project area itself. The following chart shows how project costs decrease for larger loading ratios - costs at ratios of 15 or greater are 25% less than costs for ratios of 10 and under. Also it appears that costs level-off for ratios of 15 and greater (i.e., the average cost for a loading ratio of 15 or greater is the same as for a loading ratio of 10 to 15).

Green Infrastructure Operation and Maintenance Cost

Operation and maintenance costs have been reported as well and show a wide variability. The following chart shows cost per impervious drainage area by broad type of green infrastructure, whether a subsurface or surface feature. The data indicates that surface features - those that are vegetated - cost on average more than subsurface features to maintain.

The average cost per impervious acre of $8000 equates to about $20,000 per impervious hectare. The following chart shows the variability in operation and maintenance costs according to each specific green infrastructure type. The chart shows for example higher costs for surface bumpouts and rain gardens than subsurface trenches and basins. For example, on average a bumpout costs almost twice as much as a subsurface basin.

The operation and maintenance cost appears to be approximately $20,000/$872,000 = 2.3% of capital cost. Lifecycle replacement / reconstruction of green infrastructure features, based on their deterioration over time,  would generally add to this cost and could be considered to be 1-4% of capital cost depending on the service life of the feature (i.e., features that last 25 years add 4% depreciation, and those that last 100 years add 1%).

Using these unit costs, overall lifecycle costs for Ontario-wide implementation are explored below, assuming an initial 50-year build-out period and a range of green infrastructure measures with service life durations of 25 to 100 years.

Given 852,000 urban hectares in Ontario, and assuming these are 50% impervious, the cost of green infrastructure retrofits in this province would be $370 billion dollars in capital construction cost (using $872,000 per impervious hectare) - that compares to the current Ontario stormwater infrastructure deficit of $6.8 billion. The Ontario-wide annual operation and maintenance cost for 426,000 impervious hectares would be $8.5 billion assuming $20,000 per impervious hectare - that O&M cost is over 1% of Ontario's GDP. Based on these costs, green infrastructure policies that prescribe wide-spread implementation require careful review for affordability. To recap:

Capital cost = $366 billion (using slightly lower unit cost of $860,000 per Row 12 below)
Annual O&M cost = $8.5 billion
Annual depreciation = $7.2 billion
Annual lifecycle cost (O&M + depreciation (reserve/rebuild)) = $15.8 billion

The following table summarizes the unit costs and illustrates the Ontario-wide costs that should be a cause for concern.

Ontario Green Infrastructure LID Capital, Operation and Maintenance and Lifecycle Depreciation / Reconstruction Costs - Units Costs per Philadelphia Green City, Clean Waters Pilot Program Final Report  

The follow chart illustrates the time series of costs including initial capital construction, operation and maintenance ramp-up followed by sustained operation and maintenance, reserve contributions for lifecycle asset reconstruction / rebuild according to service life (assumed 1/3 25-year, 1/3 50-year and 1/3 100-year durations), and rebuild costs (starting in year 26). It is assumed that 50 and 100-year service life assets are rebuilt over 50 a 50 year period, similar to the initial construction period.

Ontario Green Infrastructure LID Capital, Operation and Maintenance and Lifecycle Depreciation / Reconstruction Costs - 50-year initial buildout and ongoing replacement of assets beginning in year 26, funded by annual reserve.
After the initial build, the average annual operation and maintenance and depreciation costs (that are reflected in the reserve and rebuild costs) is $15.8 billion.

Some academics, including those who promote green infrastructure for amenity or other stormwater management values, have proposed green infrastructure for the purpose of flood control as well. In order to achieve flood mitigation benefits, however, widespread implementation in the sewersheds or tributaries that have flood risks is required - in that case, the costs would appear to be prohibitive to achieve quantifiable flood reduction benefits. For illustrative purposes, a York Region 100 hectare catchment has recently undergone sewer capacity upgrades at a capital cost of approximately $20M and with nominal changes in net operation and maintenance cost (larger sewers replace older ones) and a 100 year service life - implementation was over 3 years. In comparison, the green infrastructure capital costs would be in the order of $872,000 * 50% impervious * 100 hectares = $44M with additional operation and maintenance costs and lower service life durations of 25-100 years, and long term implementation (over decades) with challenges on implementation on private properties, challenges with implementation in newer tributary catchment areas with low flood risk and high existing asset value (i.e., no co-benefits of watermain replacement, etc.). Basically, the conventional flood mitigation (grey infrastructure) approach is less expensive, has a shorter implementation time and more reliably addresses the flood risk issue (i.e., green infrastructure infiltration can aggravate wastewater inflow and infiltration stresses, can adversely affect foundations, and can be unreliable in high groundwater tables areas or during saturated conditions when green infrastructure storage in ineffective).

Some further case studies and detailed assessment are required to explore where and how some green infrastructure features can contribute to Ontario urban flood risk goals in a technically effective, timely and cost-effective manner. Similarly, analysis is needed to evaluate the strategic role of green infrastructure for achieving other stormwater management goals beyond flood risk mitigation.


How do Philadelphia GSI / green infrastructure costs compare to those of other jurisdictions? One can compare unit costs of $872,000 per hectare for Philadelphia's 1,100 projects with those in Onondaga County, New York. Costs for various types of green infrastructure measures are summarized in a recent article:

The following chart illustrates lower unit costs with larger projects projects, similar to the Philadelphia reporting.

Green Infrastructure Unit Cost by LID (GSI) Type - Onondaga County, New York
These construction costs may be expressed as costs per area for projects. Considering projects managing 1 to 1.5 acres of impervious area the average cost per acre and hectare are summarized in the table below.

Green Infrastructure unit cost for projects managing up to 1.5 acres of impervious area  - Onondaga County, New York
 Excluding green roof projects, the average construction cost per impervious acres managed is over $368,000, or $783,000 per hectare. This cost is close to the Philadelphia cost of $872,000 per impervious hectare. Assuming 80% impervious surfaces in a catchment, the unit construction costs for project excluding green roofs in Onondaga County, New York is about $627,000 per hectare. This value is in the range of Ontario pilot projects with costs average costs of $575,000 per hectare.

The article citing Onondaga County green infrastructure costs notes that lower costs can be achieved by bundling implementation with other roadway works. In those cases costs were $320,000 per impervious hectare, or approximately $288,000 per total hectare, assuming 90% impervious coverage in those street projects.

Operation and maintenance costs for green infrastructure are summarized by CH2M as well. One observation that is similar to Philadelphia cost reporting is that vegetated systems are more costly to maintain than non-vegetated systems. The following chart summarizes costs per impervious area for various green infrastructure (LID, GSI) measures.

Green infrastructure operation and maintenance costs by type per impervious area managed.
Excluding green roof measures, a annual maintenance costs range from about $500 per impervious acre (low range for infiltration trench) to $3300 per impervious acre for tree infiltration trenches. A typical cost would be about $1500 per acre per year ($3700 per impervious hectare per year) which is 1500/368,000 = 0.4% of capital cost. This is significantly below the Philadelphia unit cost of $8000 per impervious hectare. It is also significantly below reported O&M/capital costs ratios reported in the American Society of Civil Engineers' report Cost of maintaining green infrastructure. In that report O&M costs for infiltration trenches and bioretention ranged from 5-20% and 5-7% respectively per one source (USEPA 1999 summarized by Weiss et al. 2007), and 8% for bioretention per another (Normalized UNHSC Installation and Maintenance Cost Data).

Basement Underpinning and Sewer Back-up Risks - How Lowering Basements Increases Flood Damage Potential in Canadian Cities Undergoing Intensification

Underpinned Basement - Typically Basemnent Floors Are
Lowered By 2 to 2 and a half feet (60 to 75 cm).
Many explanations for flood damage losses in Canadian cities have been identified - these include urbanization that extended new development areas and intensification within existing development areas, sometimes with limited runoff control. But another type of development can have a clear impact on basement flooding, or sewer back-up risk, and that is the lowering of basements. As property  homeowners strive to gain the most out of their living spaces in older urban areas where basement heights were often limited, underpinning of foundations can support the lowering of the floor slab to increase headroom and maximize the use of a finished basement.

But what does lowering the floor slab do to flood risks? Essentially it reduces the safety factor against sewer back-up but putting the floor and finished contents closer to the elevation of the municipal sewer in the street. In older areas, municipal sewer can be prone to surcharge during extreme rain events, such that sewage and extraneous water rise up well above normal flow levels. So lowering a basement reduces the 'freeboard', or buffer, between a home's valuables and the the level of wastewater in the municipal collection system. Typically a basement is lowered by 2 to 2 1/2 feet (i.e., 60 to 75 centimetres) to give a finished basement height of 8 feet (2.4 metres).

Toronto Open Data provides statistics on building permits and description of works that may include underpinning. As house prices have increased in Toronto, homeowners are clearly motivated to increase the amenity value of their basements, whether as second suites or elaborate "man-caves". Trends from 2001 to 2017 are shown below. Several hundred more permits were issues for projects including 'benching' but are not shown. Overall the number of lowered basements increased consistently from less than 200 properties in 2001 to over 1700 properties per year in 2017 - in total 14,000 properties were lowered (adding benching projects to underpinning projects shown).

Toronto Basement Underpinning Permits 2001 to 2017 - Lower Basements Can Result in Higher Flood Risks 

The following cross sections illustrate how lowering a basement reduces the 'freeboard' safety factor, or clearance, between the finished basement floor and the municipal sewer system.


Where sewer elevations are not favourable, ejector pumps for sewage may be proposed which would create a good hydraulic break between the municipal sewer and the lowered / underpinned basement.

In any case, where a basement is lowered, especially when the new space is then finished and filled with valuable contents, a backwater valve should be installed to reduce back-up risk. This can only be considered where downspouts and weeping tiles drains (foundation drains) are separated from the sanitary lines, to avoid flooding the home upstream of the backwater valve with local drainage.

Decrease in Southern Ontario Design Rainfall IDF Curves Matches Trends in Observed Storms - Decrease in Both Frequent and Rare Short Duration Intensities - Overall Decrease in Small Storms, Large Storms Mixed

Are Ontario Rainfall Trends a Nothing-Burger?
Read This Post and Find Out !
Previous posts reviewed trends in observed maximum series of observed rainfall, showing more decreasing trends in Southern Ontario than increasing trends (see post). That observed trend analysis is part of  Environment and Climate Change Canada's Engineering Climate Datasets, Version 2.3. Design rainfall intensities are derived from these observations to create intensity-duration-frequency (IDF) curves, by fitting a probability distribution to the observations. A sample of the change in design intensities over time was presented at the National Research Council's February 27, 2018 Workshop on adaptation to climate change impact on Urban / rural storm flooding  (see slides 9 and 10):

The sample IDF review showed no change in 2-year to 10-year return period intensities over durations of 5-minutes to 2 hours. The slide content was also featured in a previous post which includes links to the earlier 1990 datasets used in the comparison (for those who have thrown out those old 5 1/4 inch floppy disks with the 1990 data).

This post shows the change in IDF values for these Southern Ontario climate stations for all durations and all return periods. The chart below summarizes the change in IDF values for the 21 stations, each with 30 years of record or more. It shows the range in IDF change for each return period, across all durations. The changes for each station have been weighted by the duration of the climate station record, so that a station with a record of 60 years is given double the weight of a station with 30 years of record.

Ontario IDF Trends for Extreme Rainfall Climate Change Effects
Southern Ontario IDF Trends - Decreasing Frequent Storm Intensity, Mixed Infrequent Storm Intensity, Overall Decrease in Average Rainfall Intensity Values for Engineering Design. 5-Minute to 24-Hour Durations.
Looking into the details, the next chart shows the change in rainfall intensity for each duration within each return period as well.

Ontario IDF Trends for Extreme Rainfall Climate Change Effects Details
Southern Ontario IDF Trends - Decreasing Short Duration Storm Intensity (5 minutes - dark red bars), Decreasing Moderate Duration Storm Intensity (1-2 hours - green bars), Negligible Change in Long Duration Storm Intensity (12-24 hours - dark blue and purple bars).
The take-aways from the IDF update comparison :

i) small frequent storms (2-year, 5-year, 10-year return periods) used to design storm sewers, for example, are consistently smaller now than in the 1990 dataset,

ii) large infrequent storms (25-year, 50-year, 100-year return periods) used to design major drainage systems and infrastructure networks are mixed with some increases and some decreases since 1990 but no appreciable change that would affect design (any changes are less than 1%, which is negligible in engineering design),

ii) there is an overall average decrease in IDF values of 0.2 % across all return periods and durations.

Percentage IDF change values shown in the detailed chart are summarized in the following table for 5-minute, 10-minute, 15-minute, 30-minute, 1-hour, 2-hour, 6-hour, 12-hour and 24-hour durations, and for 2-year, 5-year, 10-year, 25-year, 50-year, and 100-year return periods.

Ontario IDF Update Trends in Rainfall Intensity and Frequency
Southern Ontario Rainfall IDF Trends From 1990 to Current Version 2.3 Engineering Climate Datasets (Average Values for 21 Long-Term Climate Stations Below 44 Degrees Latitude - Individual Station Percentage Changes Factored by Length of Climate Station Record).
Percentage IDF change values for 'unweighted' station changes (i.e., short records are given the same weight as long records) are summarized in the following table - same overall pattern as the record-length-weighted table above.

Ontario IDF Update Trends in Rainfall Intensity and Frequency Unweighted
Southern Ontario Rainfall IDF Trends From 1990 to Current Version 2.3 Engineering Climate Datasets (Average Values for 21 Long-Term Climate Stations Below 44 Degrees Latitude - Individual Station Percentage Changes Factored by Length of Climate Station Record).
So what are we to make of this? The media, the insurance industry, and those who are exercising their 'availability' bias instead of looking at storm statistics, have regularly reported that storms are bigger, or more frequent, or both, but the local Ontario data shows the opposite (Northern Ontario will be a different story as AMS trends were up in the north, unlike the south). The Ontario government  website is even out of step with the data.

The new Progressive Conservative government in Ontario has just renamed the Ministry of Environment and Climate Change the Ministry of the Environment, Conservation and Parks, taking out 'climate change', but the content under it has not been updated.

Ontario Ministry of the Environment, Conservation and Parks replaces former Ministry of the Environment and Climate Change. New name but content still reflects climate change effects on storms that is inconsistent with data.
If we look at rainfall trends in Southern Ontario it would seem appropriate to now de-emphasize the change in 'climate' or, regarding storms, the change in weather statistics. The current "MOECP" website reflects the earlier MOECC, and indicates that climate change has caused extreme weather issues in the province.

Ministry of the Environment and Climate Change website links extreme weather with climate change.
Specifically, the website indicated (as of July 2, 2018):
"It damages your property and raises insurance premiums:
  • the severe ice storm in December 2013 resulted in $200 million of property damage in OntarioToronto lost an estimated 20% of its tree canopy during the storm
  • Intact Financial, one of Canada's largest property insurers, is raising premiums by as much as 15-20% to deal with the added costs of weather-related property damage
  • Thunder Bay declared a state of emergency in May 2012 after being hit by a series of thunderstorms, flooding basements of homes and businesses due to overwhelmed sewer and storm water system"
While we cannot comment on ice storms, the official datasets for rain storms show no change, and therefore raised insurance premiums must be due to other factors instead of climate change. Blog readers will point to our review of  urbanization, intensification, etc. as a key cause.

KPMG has also commented in "Water Damage Risk and Canadian Property Insurance Pricing" (2014) for the Canadian Institute of Actuaries that prior to 2013, flood insurance pricing was inadequate, so the 15-20% increase by Intact Financial is just catching up to the market pricing for that service. It also reflects the higher value of contents and finishing of basements that are flooded / damaged during extreme weather.