Tim Johnson
City Of Mitcham & University Of South Australia
Introduction
In 2009 the City of Mitcham retrofitted twelve permeable pavement sections and tree planting pits along a two metre wide verge in a residential street to study the effects of pavement permeability on tree growth, soil moisture, ground movement and soil oxygen content. One research aim was to investigate whether seasonal desiccation of roots beneath permeable pavements during the region’s hot, dry summers might limit shallow root growth and prevent tree root related footpath damage. Observation of root development after five years of growth revealed that the permeable pavements are less prone to damage by shallow root growth than conventional impermeable pavements.
Roots observed in crushed rock screenings beneath the permeable pavements rarely exceeded 2 mm in diameter, with most being 1 mm in diameter or less. Fragments of dead roots of this size were found in the crushed rock, revealing their ephemeral nature. Further evidence of the seasonal nature of root development and the rapid turnover of fine roots was found at the interface between the crushed rock screenings and the soil; numerous root tracks were clearly visible on the soil surface but few roots remained at the time of observation. Biopores up to 2 mm in diameter descended into the soil from these root tracks, revealing that short-lived roots were not restricted to the pavement base screenings and soil surface, but that turnover of fine roots also occurred at some depth below the interface of the pavement base and soil. Permeable pavements on uncompacted, moderately reactive silty clay soil have demonstrated unfailing serviceability over six years since their construction.
Impermeable segmental concrete block pavements installed on a sand bedding layer over compacted subgrade served as controls. Excavation of these control pavements revealed shallow root development in the sand bedding layer immediately beneath the impermeable pavers. Being less able to penetrate the compacted base and subgrade these roots developed laterally beneath the paving, where continued growth could be expected to dislodge the blocks and create hazards.
Materials and methods
‘Ecotrihex’ (Adbri Masonry, Adelaide Brighton Ltd, Adelaide, Australia) concrete block permeable pavement surfaces were constructed on a base of angular rock screenings without fines over the site’s silty clay loam. Each 4 m long and 2 m wide pavement section contained a tree planting space measuring 600 mm x 1600 mm located centrally at the back of the kerb (Figure 1), in which was planted a Callery Pear (Pyrus calleryana ‘Chanticleer’) sapling. At six of the permeable sites the base layer screenings were 150 mm deep above a level soil interface, at the remaining six sites the soil was formed into a swale beneath the footpath, the swale then being filled with screenings to a maximum depth of 300 mm (Figure 2). The surrounding impermeable footpath pavement of ‘Villastone’ interlocking concrete blocks constructed on compacted site soil and a layer of bedding sand served as controls. Site and construction details have been previously reported (Johnson 2011, Johnson et al. 2011).
During winter 2014 a transect was excavated across three of each of the permeable pavement designs and three impermeable control pavements. Transects were located 1 m south of the Callery Pear trees and formed a right angle to the kerb. To expose the soil at these locations a ~1 m wide section of pavement surface was first removed to reveal the crushed rock screenings base layer. Screenings were removed using a custom built vacuum (G. S. Civil, Edwardstown, South Australia) which, with careful use, was able to extract the screenings whilst leaving roots larger than ~0.5 mm diameter intact.


Results
Roots Growing In Permeable Pavement Crushed Rock Base Material
Removal of the permeable pavement surfaces revealed networks of higher order roots (Esau 1965) living in the screenings, including very fine (<0.5 mm diameter), fine ( 0.5 – 2 mm diameter) and small roots (2 – 5 mm diameter) (Böhm 1979 in Zobel & Waisel, 2010). Roots were observed throughout the full depth of the base layer (Figure 3). These fine, intricately branched roots (labelled ‘A’ in Figure 3) were observed in the base screenings of all six of the permeable pavements excavated. They appeared identical to roots observed two years earlier during a preliminary investigation in July 2012. The roots appeared randomly distributed in the crushed rock matrix, with fine and small roots typically distributed through lower parts of the screenings and very fine roots located higher and into the finer bedding screenings immediately beneath the paving blocks.
Root tips appeared divided or flared (Figure 4) and often ended in nodules. In undisturbed base materials at the edges of the excavation many of the root tips were attached to the rock screenings. Though no water had been used in the vacuum process the surfaces of the screenings were notably damp. Screenings were covered with fine material, possibly quarry dust and some few sand particles which were amongst the screenings when installed; these were thoroughly wet but remained adhered to the rock surfaces. The origins of the very fine roots were traced back to fine and small roots originating beneath the geotextile (labelled ‘B’ in Figure 3). Few of the roots observed in the crushed rock screenings grew to more than 2 mm in diameter. No taper was apparent along unbranched sections of exposed fine and small roots, many of which exceeded 400 mm in length.

Points of emergence of the roots through the geotextile appeared to be located randomly; in most cases no evidence of any damage, previous penetration or lack of homogeneity in the fabric was observed which might have facilitated root penetration. Roots entering the screenings from the adjoining garden at one site however appeared to be aligned along a fold in the fabric at approximately the depth of the interface between the base and bedding screenings (Figure 5).


Roots were observed to have grown into permeable pavement base screenings through routes where soil covered the geotextile around the edge of the tree planting pit, though such occurrences were uncommon. The 45o soil batter at the edges of the planting pits resulted in a 70 mm depth of pavement base screenings beneath the tree pit’s pavement header course. It had been thought that seasonal desiccation of these screenings may eliminate the development of lateral roots at this shallow depth. Some few roots were observed to have developed over the top of the geotextile however, in locations where it had been inadvertently cut low around the tree pit near the base of the paving bricks or where soil disturbed during planting had settled above it. In these cases a ‘soil bridge’ existed above the geotextile through which roots grew from the tree pit into the pavement base screenings.
Dead And Decaying Roots In Crushed Rock Screenings
Pavement base screenings appeared free from organic matter other than roots. The vast majority of roots uncovered were alive and intact, but decaying and fragmented dead roots were also observed (Figure 6). Fragments of dead roots were typically 1 to 2 mm long (labelled ‘A’ in Figure 6), but the largest observed were up to approximately 25 mm in length (labelled ‘B’). The thickest decaying fragments were approximately 2 mm in diameter, similar to the diameter of the largest live roots encountered in the screenings.

Root development in grout between permeable pavers
Root systems of recently germinated herbaceous weeds were observed growing between the pavers (Figure 7). In all observed cases the roots descending from these plants terminated before entering the pavement’s bedding screenings. As the weeds were only at the cotyledon stage of development and their roots extended approximately 40 mm between the paving blocks it appeared likely that they may extend into base materials during the spring growth phase. Moss covered many of the joints and voids between the pavers at the time of examination. Moss rhizoids were not expected in the 5 mm – 7 mm screenings of the bedding layer and none were observed. Rhizoids were observed in a layer of what appeared primarily organic material which had built up on the top of the grout screenings to a maximum thickness of approximately 9 mm (Figure 8). Weed growth was routinely controlled by the City of Mitcham during winter and spring using knockdown herbicide.


Root Development through Geotextile Fabric
Roots were commonly observed to have penetrated the geotextile fabric from beneath, grown upward into and then divided amongst the permeable pavement base screenings. Occasionally these had developed a woody structure which was self-supporting, facilitating their observation in-situ following removal of the screenings. These roots divided at short intervals (Figure 9), reducing to the small and fine sizes described previously. Roots were also observed which had grown into the matted geotextile fabric, elongated within its fibres for distances of several centimetres (Figure 10), and then penetrated through the fabric into the pavement base layer screenings. In another instance, a root was observed to have grown into the geotextile from the soil, elongated within the fabric’s fibres for a distance of approximately 18 mm, then re-emerged from the fabric’s underside, re-entered the soil and continued to divide and flourish (Figure 11).



Root Development at The Soil/Geotextile Interface
Though many roots penetrated the geotextile it was noted during all excavations that much of the fabric had no penetrations (Figure 12). Lifting the geotextile revealed, as expected, similarly sparse root presence beneath as had been observed above, with large areas of the exposed soil surface being devoid of roots and the mottled surface being covered with networks of tracks (Figure 13). The quantity and density of tracks far exceeded the quantity and density of visible roots.
Tracks in the soil surface were reflected in lines on the underside of the geotextile. They were present in a range of sizes, with the broadest tracks reflected by the broadest lines on the geotextile; fine tracks matched correspondingly fine lines on the fabric. Visible roots aligned with the boldest lines on the fabric (Figure 14).

Most but not all of the tracks in the soil were three dimensional; they appeared incised into the surface. The largest were typically between 2 and 3 mm in width with some having similar depth. Where the imprinted lines on the geotextile were dense and obvious, the corresponding tracks in the soil were sharply defined with cleanly-cut edges and consistent depth (Figure 15). Less dense lines on the fabric reflected shallower and less well defined tracks in the soil; these appeared as though their edges had weathered and collapsed slightly. Where lines on the geotextile had minimal density the corresponding tracks in the soil surface appeared to have little or no depth and their edges sometimes appeared smudged and indistinct. The reduced density of some of the imprints on the geotextile, and the corresponding weathered edges and reduced depths of the tracks in soil surface, suggest that soil may have washed off the geotextile and down into the tracks during successive rainfall events. The varying densities of the many geotextile imprints and the corresponding states of the tracks in the soil surface suggest that successive generations of roots had grown, died and decayed and that multiple rainfall events had taken place to progressively wash the soil from the geotextile back into the tracks beneath.

The root tracks were noticeably more abundant in the lighter brown areas.
The tree planting pit is to the bottom left of the image.
At all six of the permeable pavements that were excavated the deepest, sharpest root tracks in the soil surface and the densest imprints on the geotextile occurred at the site of roots growing and visible at the soil surface. No biological activity other than root development was visible to the naked eye at any of the permeable pavement sites during the excavations; it was surprising that no invertebrate life was observed. In the absence of any other observed biological activity it seems most likely that the fading tracks and lines show the locations of previous generations of roots, with track depth and imprint density being inversely proportional to the time elapsed since the root’s decomposition.
The existence of pores in these tracks further supports this view. Multiple pores in the soil surface were clearly visible to the naked eye (Figure 16). Typically located at an end and descending from the bottom of the deeper lines some of the pores were up to 2 mm in diameter, though many were much smaller. Tracks left by decayed roots seem to be the most probable origin of such biopores.


Some tracks in the soil had no depth (C) and were visible due to colour variation only.

The mottled appearance of the soil surface (Figures 13, 14, 15 and 16) resulted from colour variation between the typical red-brown of the local clay-loam and a much paler shade of brown which had settled in slight depressions on the soil surface. The pale shade was not observed during pavement construction. The underside of the fabric was discoloured in these areas with the same shade of pale brown, but not in areas where the soil was red-brown. The top side of the geotextile was similarly discoloured in some locations where it had been in physical contact with the crushed rock screenings. It is therefore possible that these paler brown, lower lying areas are sites of deposition of minerals originating on or in the screenings. The discoloration of these pale areas by root growth and the redness of the lines resulting on the surface show that the discoloration had merely covered parts of the surface and not penetrated to any depth during the five years since pavement construction, suggesting fine particulate rather than soluble matter.
Root Development beneath Impermeable Control Pavements
Lifting the paving bricks at conventional impermeable footpath control sites in preparation for root examination in the subsoil gave an opportunity to investigate shallow root development in the sand bedding and at the surface of the compacted soil base. Lifting the paving at one site revealed a track resulting from the growth of a fine root at the interface of the paving brick underside and the top of the sand bedding layer (Figure 17). Clearly visible in the top of the undisturbed sand layer, root division and taper revealed the direction of growth was from the adjoining garden toward the road (to the bottom of Figure 17), indicating that the root was from vegetation in the adjoining garden. No trace of any root material was observed in the track or under the bricks. Vegetation adjoining the site included Kikuyu turf (Pennisetum clandestinum) abutting the verge, a 6 m tall Purple-leaved Plum (Prunus cerasifera ‘Nigra’) 6 m from the pavement and a 1 m tall Privet (Ligustrum sp.) growing 2 m from the site. Nearer the research tree a single very fine root was observed at the interface between the paving bricks and the sand bedding layer (Figure 18); the right hand branch of which tracked directly beneath the joints between paving bricks.


Excavation at another impermeable control site revealed compacted base aggregate and pieces of asphaltic concrete (bitumen) pavement material. Surprisingly given the youth of the tree, but not surprising given the nature of the growing media, shallow root development was observed 1 m south of the tree (Figure 19) following removal of the bedding sand layer. Between the compacted base and asphaltic concrete below, and the paving bricks above, the bedding sand layer provided the only suitable medium for root penetration and growth. Continued radial expansion of these roots would lead to dislodgement of the paving bricks and result in hazards, increased maintenance requirements and reduced asset life.

Excavation of a third impermeable control footpath revealed that no roots from the research tree had penetrated the bedding sand layer, but roots from a Silver Birch (Betula pendula) in the adjoining garden had (Figure 20). The base layer at this site was comprised of quarry rubble and the local silty clay, suggesting minimal excavation at the time of the original concrete block pavement’s construction in 1999. Compaction of the base layer was standard procedure at that time, which suggests why roots developed across it (Figure 21); the complete absence of roots penetrating the base material was noted. The two largest roots observed (Figure 21) were 13 mm in diameter near the footpath edge; at the time of excavation it was noted that their tips had died, bark had delaminated and decay was evident.
Discussion and Conclusions
Observation has confirmed that tree roots access porous crushed rock base materials beneath permeable pavements. Fine root growth ascending seasonally into surface soils from larger lateral roots deeper in the soil profile is normal in natural woodland and forest situations in temperate regions with well-defined wet and dry seasons; similar growth beneath engineered pavements has not, to the author’s knowledge, been previously reported. The observed abundance and wide distribution of roots throughout permeable pavement base materials suggest considerable benefit to trees. Decaying root parts up to 2 mm in diameter showed that in the crushed rock base material some roots of this size were short-lived. The absence of larger woody roots, the proximity of the observations to the trees and the five year growth period suggest that Chanticleer callery pear root growth in permeable pavement base materials is likely to be seasonal and that root death at 2 mm in diameter or less may be normal under Adelaide’s climatic conditions. The fine root development and evidence of frequent fine root turnover observed beneath the permeable pavement surfaces shows that more natural tree root growth and function can be engineered into our cities to achieve better performance of both the urban forest and the civil engineering.


Root tracks observed at the interface of the pavement base materials and the soil further support the view that shallow root growth was ephemeral in nature. Root tracks in the soil surface and imprinted on the underside of the geotextile confirm the size and location of successive generations of roots, most of which were no longer present at the time of observation.
The five year growth period and the diversity of tracks observed suggest that fine root turnover in this instance was more rapid than periods of several years observed in some natural systems (Lukac 2011, Tierney and Fahey 2002), though variation between species (Majdi et al. 2005) and root depths are known, with deeper roots typically living longer (Joslin et al. 2006).
The numerous biopores linked to root tracks beneath the permeable pavements are significant; they suggest biologically improved connection of the pavement’s stormwater detention capacity to the water storage capacity of the underlying silty clay loam. Roots are known to be able to rapidly transport water from zones of high potential, such as in permeable pavement base materials following rain, to recharge soil moisture at sites of low water potential through the process of hydraulic redistribution (Neumann and Cardon 2012, Prieto et al. 2012). When ephemeral roots decompose the resulting biopores would, for a time at least, continue to support increased water infiltration. No biopores were observed in the compacted soil beneath the conventional pavement’s sand bedding layer.
Soil compaction is highly undesirable from an arboricultural perspective as it inhibits root penetration (Watson 2011), reduces soil aeration (Watson & Kelsey 2006) and hampers water infiltration. Roots growing in bedding sand between impermeable pavement surfaces and compacted soil and base materials can therefore, if they survive, be expected to thicken with time and damage the pavement (Lesser 2001). Considerable care was taken during construction to avoid compaction beneath the permeable pavements; the pre-existing footpath’s compacted base was extracted with an excavator to the depth required for the rock screenings base materials. The successful function of the research pavements since 2009, built on uncompacted subgrade, demonstrates their serviceability and suggests that soil compaction is unnecessary for permeable pavements bearing light loads over moderately reactive silty clay loams. Observations further suggest that compaction of permeable pavement subgrade is detrimental as it reduces surface drainage and soil infiltration capacities. Optimising the bio-infiltration of stormwater by and in support of urban vegetation for human health, climate change adaptation, broader environmental and drainage purposes, requires that prolific root growth be encouraged in appropriate locations, so soil compaction in such areas should be minimised.
The observed interaction of trees and permeable pavements and current financial, environmental and societal pressures necessitate further investigation of these synergies and the opportunities they present. In semi-arid and arid zones permeable paving built over uncompacted soils will reduce the incidence of root-related pavement damage and tripping hazards. Tree and pavement life cycles will be extended. Risk, injury and related claims will be reduced. These facts will see permeable paving become the standard for footpaths and other light load applications across semi-arid and arid regions. With increasing understanding of the benefits and need for urban vegetation and in-situ stormwater management, few reasons will justify the increased life- cycle cost, water management consequences and risks associated with today’s standard impermeable concrete block paving into the future. Improved infiltration beneath permeable pavements, through tree roots and biopores taking the water from the soil surface where it may be problematic to deeper in the profile where it is beneficial, will see trees designed into pavements to improve their hydraulic performance. More research is needed, but the potential for trees and permeable pavements to be mutually beneficial has been clearly demonstrated. The widespread use of permeable footpath paving across Adelaide and other semi-arid and arid areas is overdue; it is needed to retain stormwater in-situ where it will be of greatest human and environmental benefit. The potential of permeable paving to prolong pavement serviceability and tree life, to address flooding and stormwater quality issues, to irrigate the urban forest and to reduce risk and expenditure compels us to consider its more widespread use.
Further Study
The reported observations were made opportunistically whilst conducting related research; they were restricted to wet and cool winter months of July and August when the trees were in their dormant phase. Information is not yet available regarding root occupancy of pavement base materials during other seasons. Further investigation is required to determine long term root growth characteristics. It is hoped that these observations will encourage further investigations to address such questions as:
- What is the fine root turnover period at the soil/base interface? When does root growth begin and cease and what are the seasonal growth rates?
- How do water infiltration rates at the interface of base materials and screenings vary between permeable pavement sites with trees and those without? Do infiltration rates vary between tree species?
- Are mycorrhizas present in the pavement base layer? Could mycorrhizas be successfully introduced into the base and subsoil through permeable paving?
- How much water can trees extract directly from permeable pavement base materials? Might some materials serve trees and other soil biodiversity better than others?
- Silty clay loam has been shown to adequately support permeable footpath pavements without compaction. What other soil types can support similarly light loads without compaction? Can uncompacted soils and permeable pavements support loadings sufficient to enable construction of driveways, car parks and streets?
- On the research site’s moderately reactive silty clay loam the permeable pavements have not induced more or less ground movement than traditional impermeable pavements. How do highly reactive soils with low hydraulic capacities respond to permeable paving in terms of ground movement, stormwater infiltration and tree root development?
- Permeable pavements reduce the amount of rainfall which reaches the soil surface though interception by and evaporation from their base materials. What is the optimal base layer composition and depth to maximise water infiltration for plant use, minimise loading on the soil surface and to still adequately support the pavement?
Acknowledgements
This research was made possible by the generous in-kind and financial support of the:
- Elected members and staff of the City of Mitcham
- Adelaide and Mount Lofty Ranges Natural Resources Management Board
- University of South Australia
- University of Melbourne
- Government of South Australia and the Department of Environment, Water and Natural Resource
References
- Böhm W. 1979. Methods of studying root systems. Berlin: Springer-Verlag. p. 188, cited in Zobel & Waisel (2010) A plant root system architectural taxonomy: A framework for root nomenclature. Plant Biosystems, Vol 144, No. 2, June 2010, pp. 507-512
- Esau, K. 1965. Plant Anatomy. New York. Wiley.
- Johnson, T.P. 2011. Can Permeable Pavements Reduce Conflicts Between Footpaths and Tree Roots? TREENET, 12th National Street Tree Symposium, September http://treenetmedia.com/?73
- Johnson, T. P., Cameron, D. A. and Moore, G. M, 2011. Can Permeable Pavements Reduce Conflicts Between Footpaths and Tree Roots? In Public Works – Capital Solutions, Proceedings of the International Conference of the Institute of Public Works Engineering Australasia, 21-25 August, Canberra Australia.
- Joslin, J.D., Gaudinski, J.B., Torn, M.S., Riley, W.J. and P.J. Hanson. 2006. Fine-root turnover patterns and their relationship to root diameter and soil depth in a 14C-labeled hardwood forest. New Phytologist 172: 523-535. Lesser, L. M. (2001). Hardscape damage by tree roots. Journal of Arboriculture, 27, 272–276
- Lukac, M., 2012. Fine Root Turnover. Plant and Soil, 362 (1-2). pp. 357-372.
- Majdi H., Pregitzer, K., Moren, A., Nylund, J. and Agren G.I. 2005. Measuring fine root turnover in forest ecosystems. Plant and Soil 276: 1-8, pp
- Neumann, R.B., and Z.G. Cardon. 2012. The magnitude of hydraulic redistribution by plant roots: a review and synthesis of empirical and modelling studies. New Phytologist, Volume 194, Issue 2, pp. 337-352.
- Prieto, I., Armas, C., and F. Pugnaire. 2012. Water release through plant roots: new insights into its consequences at the plant and ecosystem level. New Phytologist, Volume 193, Issue 4, pp. 830-841 Tierney, G.L. and T.J. Fahey. 2002. Fine root turnover in a northern hardwood forest: a direct comparison of the radiocarbon and minirhizotron methods. Can. J. For. Res. 32: 1692-1697.
- Watson, G. 2011. Fifteen years of urban tree planting and establishment research. In Trees, people and the built environment. Proceedings of the Urban Trees Research Conference, Institute of Chartered Foresters, Birmingham UK, 13 – 14 April 2011
- Watson, G., and P. Kelsey. 2006. The impact of soil compaction on soil aeration and fine root density of Quercus palustris. Urban Forestry & Urban Greening 4 (2006) 69-74