# Google-Earth Based Visualizations for Environmental Flows and Pollutant Dispersion in Urban Areas

^{1}

^{2}

^{3}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Materials and Methods

#### 2.1. Overall Framework

#### 2.2. KML File Format

#### 2.3. Key Algorithms

#### 2.3.1. Data Generation: An Integrated CFD/CRD Approach

_{j}) or concentration of species (C

^{(n)}), with n = NO, NO

_{2}, O

_{3}, etc.); V is the volume of the numerical mesh segment; S is the surface of the cell-face of the numerical mesh segment; n

_{j}is unit vector perpendicular to the cell-face of the mesh segment; ($\overline{\varphi {u}_{j}}$) is the correlation representing interactions between fluctuating components (turbulence contributions). In the standard RANS approach, the first term on the left-hand side (LHS) of the equation is zero. The terms on the right-hand side (RHS) represent the molecular diffusion, convection, turbulent diffusion and source/sink terms, respectively. The diffusive terms are discretized by the second-order central-differencing scheme (CDS), whereas the convective terms are represented as second-order linear (LUDS) or quadratic upwind differencing (QUDS) schemes. The coupling between the velocity and pressure is calculated iteratively with the Semi-Implicit Method for Pressure Linked Equations (SIMPLE).

#### 2.3.2. Data Load

#### 2.3.3. Vector Data Visualization Algorithms

_{1}, y

_{1}, z

_{1}) and (x

_{2}, y

_{2}, z

_{2}) denote the coordinates of the two end points of the arrow respectively; (x

_{i}, y

_{i}, z

_{i}) denotes the coordinates of points 3, 4, 5; (x

_{i}’, y

_{i}’, z

_{i}’) denotes the coordinates of points 3’, 4’, 5’; (x

_{0}, y

_{0}, z

_{0}) denotes the coordinates of the cross-point between the arrow and the circle plane; r denotes the radius of the virtual circle; l denotes half of the length of the arrow; f is a constant factor that is used to locate the circle and it can decide the angle of the arrow head with r; θ and φ denote the wind direction at the control volume centre. T denotes the translation matrix, R denotes the rotation matrix, R

_{x}and R

_{y}denote rotation matrices around the x-axis and y-axis respectively, and (a, b, c) indicates the velocity at the centre of the control volume. Now, the vector head can be directly converted into a tetrahedron employing the intrinsic KML function “polygon”. Displaying the vectors for all control volumes will result in a large size KML file that Google Earth will not be able to show smoothly. The solution is to pre-select some characteristic planes of interest and then to display vectors at these locations. In addition to the vectors, streamlines and streamlets are often used to visualize the velocity field. Here we briefly discuss their implementation in the present algorithm. In the first step, the mass-less particles are released at initial locations, and the closest control volumes are identified. Then, a value of the velocity component at these locations is calculated by applying tree-linear interpolation methods for the control volume vortices. In the second step, new locations of the mass-less particles (tracers) are re-calculated by multiplying the local velocity with a characteristic time step (here we apply the 2nd order Runge-Kutta method):

_{s}, y

_{s}, z

_{s}) denotes the start point, (x

_{e}, y

_{e}, z

_{e}) indicates the end point, $\Delta t$ denotes the time step, and (v

_{sx}, v

_{sy}, v

_{sz}) indicates the velocity of the start point S. Then find the closest control volume for the end point E and calculate the velocity (v

_{ex}, v

_{ey}, v

_{ez}) by tri-linear interpolation. To complete this second step, calculate the end of the coordinates of point E again with the averaged velocity, as follows:

#### 2.3.4. Scalar Data Visualization Algorithms

_{i}

_{1}denotes the coordinates of point (1) and i indicates possible x, y or z. C

_{iA}and C

_{iB}denote the corresponding coordinates of A and B respectively. S

_{t}denotes the scalar threshold value, S

_{A}and S

_{B}indicate the scalar value of A and B, respectively.

#### 2.3.5. Transformation of Coordinates

_{0}, y

_{0}, z

_{0}) are the ENU coordinates of one given point in the computational domain; (lat

_{0}, lon

_{0}, alt

_{0}) are the GPS coordinates of the given point; R denotes the averaged radius of the earth.

## 3. Results and Discussion

_{2}are defined). The GPS coordinates of the left bottom corner of the simulated domain are (51.928640, 4.461468, 0) (in decimal degrees), where the “0” indicates the ground level in the z-direction. The buildings of selected neighborhood are generated from the digital elevation map (DEM) dataset obtained from the city officials.

_{2}, O

_{3}and Reactive Organic Compounds (ROC). The environmental fluid mechanics is simulated by applying a RANS approach to obtain velocity, pressure and turbulence variables (turbulence kinetic energy and its dissipation rate). The locally refined hexagonal numerical mesh is used with typical control volume sizes of 3.5 × 3.5 × 0.25 m in the proximity of buildings, up to 40 × 40 × 40 m for regions above the urban canopy, in the x-, y- and z-coordinate directions, respectively. Note that the z-direction indicates the vertical coordinate. Approximately 5.4 × 10

^{6}control volumes are used for the entire simulation domain, and about 1800 obstacles are used to represent the buildings. The simulated CFD domain with imposed boundary conditions and numerical mesh used is depicted in Figure 6. At the inlet, approaching wind profile and intensity of the turbulence kinetic energy and its dissipation rate are defined from available meteorological measurements. The simulated scenario includes a light breeze wind conditions (~2 m/s) blowing from the West, which was based on the wind-rose map for that area for the selected time period. The symmetry boundary conditions are applied to the side and top boundaries.

_{2}and O

_{3}are shown in z = 2 m plane, Figure 9. It can be seen that the maximum concentrations of the NO

_{2}are observed in the proximity of the roads, which is expected since the NO

_{2}is a direct product of traffic emission (Figure 9a). It is also important to see that the local concentrations of NO

_{2}vary significantly along different sides of streets due to the imposed wind conditions. In contrast to the NO

_{2}, the ozone O

_{3}levels show a significantly different behavior (Figure 9b). This is the result of the complex convection-turbulent diffusion-chemical reaction mechanisms, which are simulated within the CFD/CRD model. This additionally stresses a potential of such advanced mathematical models to map various pollution scenarios based on different meteorological (wind intensity and direction) and traffic (low, moderate or high intensity) conditions.

_{2}(with levels of 0.5 ppm and 0.1 ppm, respectively) and O

_{3}(with levels of 0.15 ppm and 0.1 ppm, respectively) are shown in Figure 10. Note that this way of presenting of results can be very useful to identify regions with increased risks for population suffering from the chronic respiratory diseases (e.g., asthma, chronic obstructive pulmonary disease (COPD), etc.) by a simple pre-specification of critical concentration thresholds for individual chemical species (e.g., NO

_{2}, NO, O

_{3}, SO

_{x}, particulate meter, etc.). In present work, in addition to isolines and isotherms, we also propose a simple method to visualize the local concentrations as discrete volume objects, as demonstrated in Figure 11. Here, size and color of the object define the concentration level at pre-defined locations. This method is simple for understanding and can be easily interpreted as a kind of the virtual probes/sensors at given locations. The pre-defined critical concentration thresholds can be re-scaled to provide an easy navigation through the data. For example, the red objects indicate increased and potentially harmful levels, the yellow objects indicate intermediate pollution levels, and finally, the blue objects indicate safe concentrations. This concept was used in showing the distributions of O

_{3}(Figure 11a,b) and of NO

_{2}(Figure 11c,d). Please note that the density of the virtual probes/monitors distribution can be easily adjusted for specific regions (e.g., along streets or in the proximity of particular buildings or crossroads, parks, hospitals, schools, etc.).

## 4. Conclusions

## Supplementary Materials

## Acknowledgments

## Author Contributions

## Conflicts of Interest

## References

- Golden, J.S. The Built Environment Induced Urban Heat Island Effect in Rapidly Urbanizing Arid Regions—A Sustainable Urban Engineering Complexity. J. Integr. Environ. Sci.
**2004**, 1, 321–349. [Google Scholar] [CrossRef] - Moussiopoulos, N.; Papalexiou, S.; Sahm, P. Wind flow and photochemical air pollution in Thessaloniki, Greece. Environ. Model. Softw.
**2006**, 21, 1741–1751. [Google Scholar] [CrossRef] - Yang, Y.; Shao, Y.N. Numerical simulations of flow and pollution dispersion in urban atmospheric boundary layers. Environ. Model. Softw.
**2008**, 23, 906–921. [Google Scholar] [CrossRef] - Kenjeres, S.; de Wildt, S.; Busking, T. Capturing Transient Effects in Turbulent Flows over Complex Urban Areas with Passive Pollutants. Int. J. Heat Fluid Flow
**2015**, 51, 120–137. [Google Scholar] [CrossRef] - Kenjeres, S.; ter Kuile, B. Modeling and simulations of turbulent flows in urban areas with vegetation. J. Wind Eng. Ind. Aerodyn.
**2013**, 123, 43–55. [Google Scholar] [CrossRef] - Hanjalic, K.; Kenjeres, S. Some Developments in Turbulence Modelling for Wind and Environmental Engineering. J. Wind Eng. Ind. Aerodyn.
**2008**, 96, 1537–1570. [Google Scholar] [CrossRef] - Kenjeres, S.; Hanjalic, K. Tackling complex turbulent flows with transient RANS. Fluid. Dyn. Res.
**2009**, 41, 1–32. [Google Scholar] [CrossRef] - Ermak, D.L. User’s Manual for SLAB: An Atmospheric Dispersion Model for Denser-Than-Air Release; Lawrence Livermore National Laboratory: Livermore, CA, USA, 1990; p. 144.
- Britter, R.E.; McQuaid, J. Workbook on the Dispersion of Dense Gases; Research and Laboratory Services Division HSE: Sheffield, UK, 1998; p. 168.
- Zhou, Y.; Levy, J.I.; Hammitt, J.K.; Evans, J.S. Estimating population exposure to power plant emissions using CALPUFF: A case study in Beijing, China. Atmos. Environ.
**2003**, 37, 815–826. [Google Scholar] [CrossRef] - Stocker, J.; Hood, C.; Carruthers, D.; McHugh, C. ADMS-Urban: Developments in modeling dispersion from the city scale to the local scale. Int. J. Environ. Pollut.
**2012**, 50, 308–316. [Google Scholar] [CrossRef] - Liu, D.M.; Wei, J.M.; Kong, Z.; Xu, Z.; Tian, X. Simulation of Diffusion Trend of Unexpected Gas Leak Accident based on Google Earth. Procedia Eng.
**2012**, 45, 373–377. [Google Scholar] [CrossRef] - Chu, Y.Y.; Feng, X.H.; Zhang, Q.X.; Liang, D. Simulation of Leaking Gas Diffusion in an Urban Using 3D Map Sketch up. In Proceedings of the 2014 7th International Conference on Intelligent Computation Technology and Automation (ICICTA), Changsha, China, 25–26 October 2015; pp. 650–654.
- Chen, A.; Leptoukh, G.; Kempler, S.; Lynnes, C.; Savtchenko, A.; Nadeau, D.; Farley, J. Visualization of A-train vertical profiles using Google Earth. Comput. Geosci.
**2009**, 35, 419–427. [Google Scholar] [CrossRef] - Yamagishi, Y.; Yanaka, H.; Suzuki, K.; Tsuboi, S.; Isse, T.; Obayashri, M.; Tamura, H.; Nagao, H. Visualization of geoscience data on Google Earth: Development of a data converter system for seismic tomographic models. Comput. Geosci.
**2010**, 36, 373–382. [Google Scholar] [CrossRef] - Wang, Y.; Huynh, G.; Williamson, C. Integration of Google Maps/Earth with microscale meteorology models and data visualization. Comput. Geosci.
**2013**, 61, 23–31. [Google Scholar] [CrossRef] - Zhu, L.F.; Kan, W.S.; Zhang, Y.; Sun, J.Z. Visualizing the Structure of the Earth’s Lithosphere on the Google Earth Virtual-Globe Platform. Int. J. Geogr. Inf. Sci.
**2016**, 5, 26. [Google Scholar] [CrossRef] - Le, Y.U.; Gong, P. Google Earth as a virtual globe tool for Earth science applications at the global scale: Progress and perspectives. Int. J. Remote Sens.
**2012**, 33, 3966–3986. [Google Scholar] - Akbar, M.; Aliabadi, S.; Patel, R.; Watts, M. A fully automated and integrated multi-scale forecasting scheme for emergency preparedness. Environ. Modell. Softw.
**2013**, 39, 24–38. [Google Scholar] [CrossRef] - Jacobson, A.; Dhanota, J.; Godfrey, J.; Jacobson, H.; Rossman, Z.; Stanish, A.; Walker, H.; Riggio, J. A novel approach to mapping land conversion using Google Earth with an application to East Africa. Environ. Modell. Softw.
**2015**, 72, 1–9. [Google Scholar] [CrossRef] - KML Reference. Available online: https://developers.google.com/kml/documentation/kml-reference (accessed on 26 July 2016).
- Muilwijk, C.; Schrijvers, P.J.C.; Wuerz, S.; Kenjeres, S. Simulations of photochemical smog formation in complex urban areas. Atmos. Environ.
**2016**, 147, 470–484. [Google Scholar] [CrossRef] - Doi, A.; Akio Koide, A. An efficient method of triangulating Equi-Valued surfaces by using Tetrahedral cells. IEICE Trans. Inf. Syst.
**1991**, 1, 214–224. [Google Scholar]

**Figure 1.**The flow chart of the developed algorithm with the most important visualization techniques used for showing scalar and vector data formats, which are then converted to the GPS coordinate system, and through generation of the Keyhole Markup Language KML objects, finally, displayed in the Google Earth.

**Figure 2.**The sketch illustrating the major steps in constructing three-dimensional arrows for representation of the velocity vectors.

**Figure 3.**The underlying mesh that connects the centers of the control volumes (

**left**) and a zoom-in (

**right**) showing interpolation within the control volume (defined with vortices A-B-C-D) used to construct the characteristic isolines (red lines).

**Figure 4.**Division of a cuboidal control volume into 6 tetrahedrons required in modified marching tetrahedron approach used for plotting of the isosurfaces or isoframes of scalar variables.

**Figure 5.**The set-up of the simulated case shown in the Google Earth: region (The Netherlands) scale; city (Rotterdam) scale and neighborhood (the “Oude Noorden”) scale including computational mesh and obstacles representing parts of buildings. In total, ~1800 obstacles are included in simulations. The traffic emission sources (including the three major roads) are marked in red. The region within the yellow box is geometrically reconstructed and simulated in details.

**Figure 6.**The simulated Computational Fluid Dynamics domain with imposed boundary conditions shown in the stand-alone graphical environment (Tecplot). The INLET boundary conditions are imposed to mimic selected meteorological conditions that include the wind intensity, its direction, and intensity of wind fluctuations. The OUTLET boundary assumes a zero gradient condition (in the wind direction) for all variables. The SYMMETRY boundaries impose that there is not flow perpendicular to the boundary. The red segments indicate the main traffic emission sources (roads). The numerical mesh used is also shown in characteristic planes (in total, 5.4 × 10

^{6}control volumes are used for computations). The yellow box indicates the extracted sub-domain exported for Google Earth visualizations.

**Figure 7.**Vector representation of the wind flow in characteristic horizontal (close to the ground) and vertical planes: (

**a**) overview for the entire domain at z = 1 m and y = 454 m planes; (

**b**) zoom-in. The street-canyon effect (in the horizontal plane) and wakes generated behind taller buildings (in the vertical plane) can be easily observed. Note: a length of 12 m represents the reference vector of 2 m/s.

**Figure 8.**(

**a**) Streamlines and (

**b**) streamlets portraying the details of the wind patterns and their changes at various locations due to the presence of the buildings. Similarly to the previous figure, the red segments indicate strong wind flow regions, whereas the blue sections indicate the low wind intensity regions.

**Figure 9.**The mapping of pollution by isolines: (

**a**) concentration contours of NO

_{2}(indicating the locations with high emission sources, i.e., traffic) and (

**b**) concentration contours of O

_{3}(indicating locations with enhanced ozone distribution as result of the chemical reactions)—both in the horizontal plane at pedestrian level (z = 2 m).

**Figure 10.**Visualization of characteristic pollution fronts by selecting some critical thresholds of species and connecting them as 3D isosurfaces of (

**a**) NO

_{2}and (

**b**) O

_{3}.

**Figure 11.**The virtual concentration sensors generated as the discrete volume objects, which size and color are proportional to the concentrations of species at given locations. Concentrations of: O

_{3}(

**a**,

**b**); NO

_{2}(

**c**,

**d**) across the whole domain or for particular streets, respectively.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ( http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Liu, D.; Kenjeres, S. Google-Earth Based Visualizations for Environmental Flows and Pollutant Dispersion in Urban Areas. *Int. J. Environ. Res. Public Health* **2017**, *14*, 247.
https://doi.org/10.3390/ijerph14030247

**AMA Style**

Liu D, Kenjeres S. Google-Earth Based Visualizations for Environmental Flows and Pollutant Dispersion in Urban Areas. *International Journal of Environmental Research and Public Health*. 2017; 14(3):247.
https://doi.org/10.3390/ijerph14030247

**Chicago/Turabian Style**

Liu, Daoming, and Sasa Kenjeres. 2017. "Google-Earth Based Visualizations for Environmental Flows and Pollutant Dispersion in Urban Areas" *International Journal of Environmental Research and Public Health* 14, no. 3: 247.
https://doi.org/10.3390/ijerph14030247