Using SAR Satellite Imagery for Potential Green Roof Retrofitting

for Flood Mitigation in Urban Environment

Mirka Mobilia

and Antonia Longobardi

Department of Civil Engineering, University of Salerno, Via Giovanni Paolo II, 132, Fisciano (SA), Italy

Keywords: Green Roof, SAR Images, SWMM, Sarno River Basin.

Abstract: Green roofs (GRs) represent a valid tool to mitigate the negative effects of urban floods. The aim of the present

work is to test the hydrological behaviour of GRs at basin scale using Storm Water Management Model

(SWMM), with an application to Preturo municipality located within the Sarno river basin (southern Italy)

which during the last two decades was interested by the occurrence of many hazardous flash floods. The

analysis of two sets of satellite images acquired between 1995 and 2016 by SAR sensors, showed a correlation

between the increase of soil sealing and the occurrences of urban flooding during the same period. This finding

suggests that the GR extensive adoption could contribute to a successful stormwater management. The

suitability for GR retrofit depends on a number of criteria. In Preturo municipality, the fulfilment of these

criteria was investigated using satellite images from Google Earth. The GRs retrofit potential of the studied

area amounts to 7% of the total surface. The hydrological behaviour of the GR retrofit scenario was compared

to the reference one which considers the actual land cover. The GR scenario better performs than the

traditional one with a reduction of the runoff volume and peak flow of respectively 3.5% and 18.9% and an

increase of the delay time of 8.2%.

1 INTRODUCTION

The most direct impact of the rapid population growth

is the demand for new buildings, which causes the

increase of urban areas all over the world. The

uncontrolled soil sealing leads to increasingly severe

and frequent urban flooding events (Mobilia et al.,

2015, Longobardi et al., 2016, Khramtsova et al.

2020). The low impact development (LID) strategies

consist of a number of practices, which mimic the

site’s predevelopment hydrologic functions so as to

mitigate the hazardous hydrological events (Chui et

al., 2016). Among these, the green roofs (GR) are

better able to minimize, detain and retain the extra-

runoff (Sartor et al., 2018, Longobardi et al., 2019,

Mobilia et al., 2017). Over the years, modelling

results have confirmed the role of GRs in restoring

the natural water regime by reducing the runoff

production. Some of the approaches used in previous

studies to reproduce the GRs hydrological behavior,

involved the use of water balance model (Starry et al.,

2016, Mobilia and Longobardi, 2020a, Mobilia and

Longobardi, 2017), NASH cascade model

https://orcid.org/0000-0001-7018-3592

https://orcid.org/0000-0002-1575-0782

(Krasnogorskaya et al., 2019), Hydrus model

(Mobilia and Longobardi, 2020b) and the rainfall-

runoff simulation model by Environmental Protection

Agency (EPA), SWMM (Storm Water Management

Model) (Haowen et al. 2020). In particular, the last

one has proved to be very accurate and user-friendly

(Wanniarachchi, 2012). Several studies concerning

GR hydrological modelling at basin scale have

demonstrated that, if widely implemented, these tools

allow to significantly reduce the negative impact of

urban flooding by decreasing the runoff volume even

up to 25% and the peak discharge to 36% (Versini at

al., 2015). Unfortunately, the extensive adoption of

vegetated cover in an urban basin is prevented by

some barriers, indeed, several criteria should be

considered when determining whether a roof is

suitable for retrofitting including the roof slope,

number of stories, orientation of the roof, number of

site boundaries (Wilkinson and Reed, 2009). In this

context, satellite imagery could be used to identify the

buildings with the potential for GR retrofit. The

present research aims at investigating the ability of

GRs, at large scale, to reduce the hydraulic load on

Mobilia, M. and Longobardi, A.

Using SAR Satellite Imagery for Potential Green Roof Retroﬁtting for Flood Mitigation in Urban Environment.

DOI: 10.5220/0010378700590066

In Proceedings of the 7th International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2021), pages 59-66

ISBN: 978-989-758-503-6

urban drainage system in order to mitigate urban

flooding events. Two scenarios have been simulated,

using SWMM and compared in terms of runoff

volume, peak flow and delay time during the same

rainfall event. The first scenario considers the

existing land cover while the second one concerns a

conversion of the traditional roofs to GRs in respect

of the building attributes for GR retrofit. The use of

high-resolution satellite imagery from Google Earth

has allowed to identify the buildings suitable to host

a retrofitted GR. The case study is Preturo

municipality in Southern Italy. The city has been

selected as located within Sarno river basin which is

frequently affected by urban flooding and the massive

increase in the occurrence of flash floods mainly took

place between 1995 and 2016. It has been shown that,

such an increase cannot be directly linked to climate

change (Califano et al., 2015) but it could be related

to the uncontrolled expansion of soil sealing in the

same period. In the last case, the GR retrofit of

existing roofs within the considered area could lead

to a successful stormwater management. The analysis

of the variation of impervious area has been performed

using two set of SAR (Synthetic Aperture Radar)

images: 1995s images provided by ERS-1 sensor and

2016s images supplied by COSMO-SkyMed sensor.

The obtained temporal coherence maps have been

treated with the color thresholding method for image

segmentation so as to isolate urban pixels and detect

the change in time of the paved surfaces.

2 MATERIAL AND METHODS

SWMM model has been used to predict the

hydrological response of the drainage network

resulting from two GR conversion scenarios within

Preturo municipality. The first scenario is the actual

one while the second scenario refers to the greening

conversion of buildings which meet predefined

criteria. The adoption of GRs for the retrofitting of

existing roofs in order to mitigate urban flooding

events, confirms to be an appropriate solution mainly

if within the considered area, a rapid and uncontrolled

urbanization has occurred during the studied period.

The analysis of land cover change during the last two

decades (from 1995 to 2016), has been performed by

means of the elaboration of SAR images.

2.1 Case Study

The case study is Preturo municipality in Campania

region (Southern Italy). It is a poorly populated and

urbanized area with 1768 inhabitants for a surface of

about 28 hectares. The altitude of the municipality is

about 190 m above sea level. According to Koppen

classification, the site is characterized by

Mediterranean climate (Csa) with hot, dry summers

and cool, wet winters with the highest percentage of

rain in the year. The temperature here averages 15.2°

C while the rainfall in a year, is about 845 mm. The

city has been selected as all the details concerning its

urban drainage system are available and furthermore,

it is located within Sarno river basin (Figure 1) which

is a risk-prone area where a large number of flooding

events occurred during the last decades.

Figure 1: Preturo municipality as part of Sarno basin.

Indeed, during the last two decades, the Sarno river

basin has experienced a massive increase in the

occurrence of flash floods (Figure 2)

Figure 2: Floods occurred between 1995 and 2016 in Sarno

river basin.

It has been argued that, the increase in the

frequency of flooding events between 1995 and 2016

in Sarno basin, is not linked to climate change

GISTAM 2021 - 7th International Conference on Geographical Information Systems Theory, Applications and Management

(Califano et al. 2015) but it could be related to the

rapid urbanization of the area during the same period.

If so, the GR retrofit of existing roofs would result

the most appropriate choice for urban stormwater

management.

2.2 SAR Images Elaboration

Remote sensing imagery has been used to detect the

land use/cover change transitions between 1995 and

2016 in the Sarno river basin, in order to identify a

correlation between the increase of paved surfaces

and of the hydrological damaging events. The process

here proposed for the elaboration of SAR images, can

be organized in three major blocks (Figure 3).

Figure 3: Processing algorithm.

The first step was the data download. Two sets of

archive satellite images have been used (Table 1). The

first set of images consists of 6 images acquired

between March and December 1995 by the ERS-1

satellite and distributed by the European Space

Agency (ESA), the second dataset has been acquired

by the COSMO-SkyMed sensor between May and

December 2016, and consists of 7 images distributed

by the Italian Space Agency (ASI).

Table 1: Acquisition date of SAR images.

ERS1 COSMO-SkyMed

Acquisition date

24th March 1995 3rd May 2016

8th July 1995 20th June 2016

12th August 1995 6th July 2016

21st October 1995 8th September 2016

25th November 1995 24th September 2016

30th December 1995 27th November 2016

- 13th December 2016

The pre-processing block includes data

coregistration and the estimation of the

interferometric coherence. Coregistration is the

process of alignment of the images with the reference

one, which is usually the first acquired. In the present

study, the reference images for the ERS1 and

COSMO-SkyMed time series have been acquired,

respectively, on March 1995 and May 2016.

The pre-processing block also includes the

interferometric coherence estimation. It is a measure

of the stability of a target with respect to the phase of

the complex signal. Considering two co-registered

images, it is computed through the following relation:

() ()

)()( xSxS

xSxS

⋅Σ

(1)

Where S

and S

are respectively the complex image

values for the first and the second image while *

represents the complex conjugation operation. The

reference image used in this step is the reference

image of the coregistration phase. The coherence

ranges between 0 and 1 respectively referring to the

natural land cover and the built-up area.

The feature extraction block includes the temporal

mean of the coherence consisting of averaging per

year the multi-temporal SAR images of coherence in

order to reduce the speckle but not the spatial

resolution. The second activity of the block is the

coherence threshold assessment carried out using

Otsu’s method allowing to convert a grey level image

to monochrome (or binary) image where the white

pixels represent the impervious surfaces while the

black one the previous ones according to the

following equation:

𝜎







𝑡



=𝑊





𝑡



⋅𝜎







𝑡



+𝑊





𝑡



⋅𝜎







𝑡



(2)

The best threshold value t* corresponds to the

minimum within class variance σ

. W

and W

are the

weights of the foreground and background classes of

pixels given as:

𝑊





𝑡



∑

𝑃



𝑖







;𝑊





𝑡



∑

𝑃



𝑖







(3)

Where:

𝑃



𝑖



𝑛



𝑛

(4)

The symbol i represents the gray-level and n

the

quantity of the pixels with the specified gray-level, n

is the general number of pixels in the image, I is the

maximum pixel value. σ

and σ

are the variances of

the

two classes:

𝜎





∑































with 𝜇



∑

∙

















(5)

𝜎





∑































with 𝜇



∑

∙

















(6)

Where μ

and μ

are the mean of the two classes.

Using SAR Satellite Imagery for Potential Green Roof Retroﬁtting for Flood Mitigation in Urban Environment

Finally, the processing of the two sets of satellite

images from ERS-1 and COSMO-SkyMed sensors has

allowed to detect the variation in the sealing surfaces

during the considered period and to find a possible

correlation between it and the increasing occurrence of

hazardous events within the studied area. If a link

exists the implementation of green infrastructures, with

the aim of reducing the risk associated to severe events

in urban area, can be investigated. The implementation

of these infrastructures is closely related to the

potential for retrofitting of existing roofs investigated

as explained in the following chapters.

2.3 Building Attributes for Green

Roofs

The hydrologic effect of two GR conversion

scenarios, at the city scale, has been tested using

SWMM. The first scenario considers the existing land

cover where no GRs are installed while the second

one benefits from the greening conversion of some

traditional roofs of the city. Sustainability for GR

retrofit has been detected using Google Earth and it

mainly depends on four criteria (Wilkinson and Reed,

2009) that are:

 Roof slope;

The greening should be applied to roof with a

minimum slope of 2% and a maximum slope of 45%.

Roofs with a slope less than 2% require additional

drainage measures in order to avoid waterlogging in

the vegetation support course. On the other hand, a

sloped roof retains less water and structural and

vegetation problems could occur like the slip of the

plant layer.

 Number of stories;

Taller buildings could partially or totally

overshadow the adjacent smaller ones and the shadow

could affects negatively the grow of the plants and

reduces evapotranspiration fluxes. Because the height

of each building is unknown, the number of stories

has allowed to discern the taller and the smaller

buildings.

 Orientation of the roof;

In general, the sunlight contributes to the welfare

of the vegetation and in the northern hemisphere, the

exposure to direct sun is higher for south-facing

buildings.

 Number of site boundaries;

If a building is attached to others on four sides,

during the construction of GR, the access for

machinery and delivery or storage of materials

tend to be difficult just like the subsequent access

for maintenance. In view of this, the buildings

bounded on four sides should be discarded.

2.4 Model Overview, Setup and

Implementation

The EPA Storm Water Management Model (SWMM)

has been selected as the approach for studying the

runoff production in both the considered scenarios.

SWMM is a dynamic hydrology-hydraulic simulation

model originally developed for urban areas. SWMM

consists of several blocks for analysis of different

processes: the Runoff block which performs

hydrologic simulation of runoff production, the

Transport and Extended Transport block for the

routing of this runoff, the Storage/Treatment block

which characterizes the effects of control devices

upon flow and quality and elementary cost

computations and the receive block which considers

the mix of the produced runoff in a receiving water

body. SWMM is equipped with a LID module used to

model various stormwater management devices

including GRs. SWMM solves the complete form of

Saint-Venant equations over the drainage network.

The Saint-Venant equations for conservation of mass

and momentum can be respectively express as

follows:

𝛿𝑄/𝛿𝑡 + 𝛿𝐴/𝛿𝑥 = 0

(7)















⁄





+𝑔𝐴





+𝑔𝐴𝑆



+𝑔𝐴ℎ



(8)

where A is the cross-sectional area, t is the time, Q is

flow rate, x is distance, H is hydraulic head of water

in the conduit, g is the gravity, h

represents the local

energy loss per unit length of conduit; and Sf is the

friction slope. SWMM solves the Saint-Venant

equations using an explicit finite difference method

and successive approximation. The continuity

equation (Eq. 7) states that in any control volume of

water, the net change of mass caused by the inflow

and outflow equals the net rate of change of mass in

the same control volume. The momentum equation

(Eq. 8) asserts that the rate of change of momentum

in the control volume of water equals all the external

forces acting on the same control volume. Preturo’s

sewer network including the junction nodes, the

twelve sub-catchments (from SUB1 to SUB12), the

conduits can be sketched in SWMM as in Figure 4a.

The model has run with a rainfall input referring to

the event occurred on 13/09/2012 and recorded by the

local rain gauge. The rainfall event has been selected

as it caused severe flash flood in the considered area.

It lasted about 17 hours with a cumulative rainfall of

56.4 mm and a return period of 8 years. The temporal

distribution of the event is shown in Figure 4b.

GISTAM 2021 - 7th International Conference on Geographical Information Systems Theory, Applications and Management

Figure 4: a) Urban drainage system of Preturo in SWMM,

b) Rainfall input.

The GR used for the simulation is an experimental

roof located in the campus of university of Salerno. It

is of extensive type and it is made up of three layers

for a total thickness of 15 cm: the vegetation layer

where the plants grow, 10 cm deep support substrate

which ensures a suitable growing environment for the

vegetation, and a water storage layer of 5 cm where

the water can be retained.

The support layer is mainly composed of peat

while the storage layer is made up of expanded clay

(Figure 5).

Figure 5: GR structure.

More details about the characteristics of the

considered GR and the input parameter set of the GR

module used for running the model are available in

(Mobilia, 2018).

Due to the lack of flow measurements within the

basin, the calibration procedure of the model has not

been performed, anyway, since several rainfall/runoff

events have been recorded at the experimental roof,

the calibration of the LID module has been carried out

instead. For more details about the observed events,

the calibration procedure and parameters please

consult Mobilia and Longobardi (2020c) and Mobilia

et al. (2020).

2.5 The Performance Analysis

The event-based performance is assessed through

three indices: the percentage of reduction in runoff

volume (ΔRV), in peak flow (ΔPF) and increase in

the delay time (ΔDT). The indices have been

estimated as follows:

𝛥𝑅



𝑅

,

𝑅

,

𝑅

,

⋅ 100

(9)

𝛥𝑃





,



,



,

⋅ 100

(10)

𝛥𝐷





,



,



,

⋅ 100

(11)

Where R

V,0

, P

F,0

, D

T,0

are respectively the runoff

volume (m

), the peak flow (m) and the delay time

(min) referred to the baseline scenario, while R

V,1

F,1

, D

T,1

are the same parameters but referred to the

greening scenario. The indices have been calculated

with respect to the outlet sections of each sub-

catchment and of the whole basin.

3 RESULTS

Remote sensing imagery has been used to detect the

land use/cover change transitions between 1995 and

2016 in the Sarno river basin. The results of SAR

images elaboration show that the build-up area moved

from about 7% to about 12% between 1995 and 2016,

so Sarno watershed has experienced a rapid

urbanization during the last two decades (Figure 6).

This finding suggests that the increase in the

occurrence of flooding events within Sarno basin

could be attributable to the overbuilding of the

surface, therefore, the retrofitting of buildings with

Using SAR Satellite Imagery for Potential Green Roof Retroﬁtting for Flood Mitigation in Urban Environment

GRs may represent a valid solution to mitigate the

negative effects of hydrological damaging events.

The hypothesis that GRs, applied at large scale,

could be effective in managing the urban stormwater

management in an overbuilding area, is hereinafter

tested with the use of SWMM.

Figure 6: The change in built-up area between 1995 and

2016.

The hydrological behavior of the drainage

network in two scenarios has been analyzed. The first

scenario is the baseline one while the second scenario

involves the widespread GR implementation within

the area according to four factors affecting the

potential to retrofit existing roofs which are: roof

slope, number of stories, orientation of the roof,

number of site boundaries. The identification of the

suitable buildings has been visually performed by

means of the images provided by Google Earth. With

respect of the roof slope, all the existing buildings

meet the requirement. As regards the criteria of

number of stories, the buildings with such a height as

not to be overshadowed by the nearby ones represents

the 7.1% of the total area of the basin (Figure 7). In

particular, the average number of stories is two.

Overall, 24 buildings are one story, 80 are two stories

and 72 are three stories. Only 1% of the total

buildings are four stories buildings.

Figure 7: Buildings selected according to the number of

stories.

The examination of site orientation revealed that

the south-facing buildings are about 10% of the total

basin area (Figure 8).

Figure 8: Buildings selected according to the orientation of

the existing roofs.

With references to criteria of number of site

boundaries (Figure 9) which aims at discarding the

buildings bounded on four sides, it can be said that

most of houses have not neighboring buildings (29%

of the total houses) so are free-standing while the 23%

of the buildings are bounded on two sides and the

21% on three sides. Only 15% of the total buildings

are bounded on one side. In conclusion, the buildings

which comply with this criteria represent the 9% of

the total basin area. Finally, the buildings which meet

all the attributes required for GR adaptation occupied

GISTAM 2021 - 7th International Conference on Geographical Information Systems Theory, Applications and Management

on average about the 7% of the area of the whole

basin. This percentage has been used to reproduce, in

SWMM, the hydrological behavior of a hypothetical

greening scenario which has been subsequently

compared to the baseline scenario where the existing

land cover of the basin is considered.

Figure 9: Buildings selected according to the number of site

boundaries.

The comparison has been performed in terms of

produced runoff volume, peak flow and delay time.

At a first visual inspection of the runoff hydrographs

observed at the outlet section of the whole basin

(Figure 10), it is possible to observe how, the runoff

volume and the peak flow are visibly lower in the case

of the GR retrofit scenario than for the baseline one.

Figure 10: Comparison of existing land use condition

hydrograph with the greening conversion hydrograph at the

catchment outlet.

For a comparison in quantitative terms, the

assessment of three indices whose values are shown

in Table 2, has been performed. A reduction of the

runoff volume and of the peak flow of respectively

3.8% and 18.9% has been detected for the whole

basin after the GR conversion of the existing

buildings while an increase of the delay time of 8.2%

can be observed. At the single sub-catchment scale,

after the GR retrofit, the runoff reduction ranges from

the minimum value of 0.8% to the maximum value of

9.4%. The GR performances in terms of peak flow

reduction and delay time increase are lower if

compared to the ones obtained for the whole basin

indeed, they reach respectively at most the 3.6% and

0.8%. The worst performances are found in the

eastern part of the basin (SUBs 2, 4, 8, 9, 12) where

the retrofitting potential is lower.

Table 2: Values of the indices.

Sub ΔRv (%) ΔPf (%) ΔDt (%)

1 5.1 2.6 0

2 1.1 0.0 0

3 8.4 1.6 0.8

4 0.9 0.0 0

5 5.6 1.0 0

6 5.3 1.2 0

7 6.2 1.8 0

8 1.9 0.0 0

9 1.2 0.0 0

10 9.4 3.2 0

11 7.2 3.6 0

12 0.8 0.0 0

Whole 3.8 18.9 8.2

4 CONCLUSIONS

The GR technology is a valid solution to mitigate the

risk of flooding in urban areas especially in the

territory interested by a rapid and uncontrolled soil

sealing. This is the case of the Sarno river basin that

between 1995 and 2016, has experienced a doubling

of the build-up area. The aim of this research is to

investigate and compare the hydrological

performance of two scenarios respectively with and

without the adoption of GRs for the retrofitting of

existing roofs in Preturo municipality which is part of

Sarno river basin. The factors impacting the potential

for the retrofit of an existing roof are the number of

stories, the orientation of the roof, the number of site

boundaries and the roof slope which have been

verified with Google Earth. The percentage of GR

retrofit potential in the basin is 7% of the total area.

The analysis has been performed using SWMM and

showed that the widespread GR implementation at the

Using SAR Satellite Imagery for Potential Green Roof Retroﬁtting for Flood Mitigation in Urban Environment

whole basin scale reduces the peak runoff rates and

the runoff volumes respectively by up to 4% and 19

% while it increases the delay time of about 8%. In

conclusion, in poorly urbanized area, the application

of GRs at large scale, returns a good attenuation of

flooding events but the technology could be better

performing in densely populated areas. For higher

attenuation of the urban floods, the GR infrastructures

could be used in combination with different types of

LID practices. Future research directions of the

present work are twofold. From one side, the building

selection can be improved by using geospatial data

such as Lidar and photogrammetry mapping

technologies. From the other side, the identification

of the pervious/impervious surfaces can be optimized

by combining SAR and optical images. The joint use

of the both types of images allows to overcome the

limitations of the two approaches. Indeed, the optical

images have a high spatial resolution but suffer from

the problem of the cloud cover and vice versa for SAR

images.

REFERENCES

Califano, F., Mobilia, M., Longobardi, A., 2015. Heavy

Rainfall Temporal Characterization in the Peri-Urban

Solofrana River Basin, Southern Italy. Procedia

Engineering, 119, 1129-1138.

Chui, T. F. M., Liu, X., Zhan, W., 2016. Assessing cost-

effectiveness of specific LID practice designs in

response to large storm events, Journal of Hydrology,

533, 353-364.

Haowen, X., Yawen, W., Luping, W., Weilin, L., Wenqi,

Z., Hong, Z., Yan Yichen, Jun, L., 2020. Comparing

simulations of green roof hydrological processes by

SWMM and HYDRUS-1D, Water Supply, 20(1), 130-

139.

Khramtsova, L., Leonteva, M., Mobilia, M., Longobardi A.,

Nasyrova, E., Aksenov, S., 2020. Characterization of

rainfall events corresponding to forecasted conditions

of potential hydrogeological risk, Journal of Physics:

Conference Series, 1614(1),012077.

Krasnogorskaya, N., Longobardi, A., Mobilia, M.,

Khasanova, L. F., Shchelchkova, A. I., 2019.

Hydrological modeling of green roofs runoff by Nash

cascade model, The Open Civil Engineering Journal,

13(1), 163-171

Longobardi, A., D’Ambrosio, R., Mobilia, M., 2019.

Predicting stormwater retention capacity of green roofs:

An experimental study of the roles of climate, substrate

soil moisture, and drainage layer properties,

Sustainability, 11(24), 6956.

Longobardi, A., Diodato, N., Mobilia, M., 2016. Historical

storminess and hydro-geological hazard temporal

evolution in the solofrana river basin—Southern Italy,

Water, 8(9), 398.

Mobilia, M., 2018. Sustainable managment of stormwater

in a changing environment under Mediterranean

climate conditions, PH.D. dissertation.

Mobilia, M., Califano, F., Longobardi, A., 2015. Analysis

of rainfall events driving MDHEs occurred in the

Solofrana River Basin, Southern Italy, Procedia

Engineering, 119, 1139-1146.

Mobilia, M., D’Ambrosio, R., Longobardi, A., 2020.

Climate, soil moisture and drainage layer properties

impact on green roofs in a Mediterranean environment.

In Frontiers in Water-Energy-Nexus—Nature-Based

Solutions, Advanced Technologies and Best Practices

for Environmental Sustainability. Springer, Cham.

Mobilia, M., Longobardi, A., 2017. Smart stormwater

management in urban areas by roofs greening. In

ICCSA2019, International Conference on

Computational Science and Its Applications. Springer,

Cham.

Mobilia, M., Longobardi, A., 2020a. Model Details,

Parametrization, and Accuracy in Daily Scale Green

Roof Hydrological Conceptual Simulation.

Atmosphere, 11(6), 575.

Mobilia, M., Longobardi, A., 2020b. Impact of rainfall

properties on the performance of hydrological models

for green roofs simulation. Water Science and

Technology, 81(7), 1375-1387.

Mobilia, M., Longobardi, A. 2020c. Event scale modeling

of experimental green roofs runoff in a mediterranean

environment. In Frontiers in Water-Energy-Nexus—

Nature-Based Solutions, Advanced Technologies and

Best Practices for Environmental Sustainability.

Springer, Cham.

Mobilia, M., Longobardi, A., Sartor, J. F., 2017. Including

a-priori assessment of actual evapotranspiration for

green roof daily scale hydrological modelling. Water,

9(2), 72.

Sartor, J., Mobilia, M., Longobardi, A., 2018. Results and

findings from 15 years of sustainable urban storm water

management. International Journal of Safety and

Security Engineering, 8(4), 505-514.

Starry, O., Lea-Cox, J., Ristvey, A., Cohan, S., 2016.

Parameterizing a water-balance model for predicting

stormwater runoff from green roofs, Journal of

Hydrologic Engineering, 21(12), 04016046.

Versini, P. A., Ramier, D., Berthier, E., De Gouvello, B.,

2015. Assessment of the hydrological impacts of green

roof: From building scale to basin scale. Journal of

Hydrology, 524, 562-575.

Wanniarachchi, S. S., Wijesekera, N. T. S., 2012. Using

SWMM as a tool for floodplain management in

ungauged urban watershed, Engineer: Journal of the

Institution of Engineers, Sri Lanka, 45(1). 1-8

Wilkinson, S.J., Reed, R., 2009. Green roof retrofit

potential in the central business district. Property

Management, 27(5), 284-301.

GISTAM 2021 - 7th International Conference on Geographical Information Systems Theory, Applications and Management