Draft Workplan

Development of an urban flood model for Rio de Janeiro city

Augusto Getirana and Dalia Kirschbaum

April 5, 2019

Summary of Project

Rio de Janeiro has historically suffered from massive rainstorms and subsequent floods and
landslides but as yet does not have a flood modeling system. As part of the NASA-Rio partnership
and in coordination with the IPP, Fundação Rio Águas and Fundação Geo-Rio team within the
mayor’s office in Rio we have developed a plan to advance flood modeling capabilities at the city
scale. The city of Rio will provide critical in situ data that will be used together with NASA satellite
data to parameterize and test two modeling frameworks designed to look at flooding: The NASA’s
Land Information System (LIS), developed at GSFC, and LISFLOOD-FP hydrodynamic model,
developed at the University of Bristol. The outcome of this effort will be to implement the flood
modeling system in the Municipal Urban Information System (SIURB), as well as outline the
potential for this modeling capability in other cities. This project will require one full-time person
at NASA GSFC working with Drs. Getirana and Kirschbaum for up to a 2 years. The total cost for
this effort per year is estimated at $125k.

Background

In 2015, NASA signed an agreement with Rio de Janeiro, Brazil to support innovative efforts to
better understand, anticipate, and monitor natural hazards, including drought, flooding, and
landslides, in and around the city. The collaboration will leverage the unique attributes of NASA’s
tools and satellite data systems and Rio de Janeiro’s monitoring and crisis management
capabilities to improve hazard monitoring and disaster response.

Rio de Janeiro has historically suffered from massive rainstorms and subsequent floods and
landslides, all of which have cause casualties and adverse socioeconomic impacts. For example,
heavy storm, up to 90mm of rain within one hour, and wind gusts of around 110 km/h hit the city
between 6-7 February and 7-8 April this year, causing floods and landslides, and killing numerous
people. Unfortunately, such disasters occur often in the city, in particular during the rainy season
between December and March. Impacts of floods are further aggravated by the lack of a flood
monitoring and forecast system. In this sense, the goal of this project is to take advantage of
NASA tools and datasets and Rio de Janeiro ground observation system to develop a customized
urban flood model that will be further operated by Fundação Rio Águas, the local Urban Drainage
foundation. The model will be based on the NASA’s Land Information System (LIS; Kumar et al.,
2006). LIS is a software framework supporting a range of hydrological models and a variety of
datasets. The accuracy of different land surface models (LSM) available in LIS (e.g. Noah, CLSM
and VIC) will be evaluated. The total runoff simulated by LSMs will be routed through a river and
channel network. This project will involve collaborations with external partners due to the

1
absence of an urban flood model in LIS. Dr. Paul Bates (University of Bristol), main developer of
the LISFLOOD-FP model (Bates and De Roo, 2000; Bates et al., 2017), has demonstrated interest
in collaborating and making his model available for our use. LISFLOOD-FP is a 2-D hydrodynamic
model that has been widely used for high resolution urban flood modeling (e.g. Sampson et al.,
2012; Ozdemir et al., 2013). We will also explore the potential of downscaling the large-scale
HyMAP river routing scheme (Getirana et al., 2012, 2017), available in LIS, in order to evaluate
their accuracy in representing fine scale processes. We plan to use NASA datasets from Global
Precipitation Measurement (GPM) as precipitation forcing at near real time.

Data from local and regional monitoring networks within the greater Rio de Janeiro region will
also be used for model parameterization, calibration and evaluation. The Rio de Janeiro City
Government will provide the dataset composed of a network of ground-based
hydrometeorological observations and LiDAR measurements over Rio de Janeiro. However, the
dataset needs to go through an inventory and formatted before it is used in this project. Readily
available observations, as informed by local partners, are listed below:
- 33 pluviometers operated by AlertaRio;
- 26 stream gauging stations;
- vectorized river network, including channels and ponds;
- land cover and land change maps;
- map of rainwater accumulation points;
- LiDAR measurements covering two thirds of the city (the topography of the remaining area is
expected to be measured this year) (see Fig. 1).

Figure 1: LiDAR coverage and floodable areas in Rio de Janeiro (courtesy: Fundação Geo-Rio)

The project will be developed in the Hydrological Sciences Laboratory at NASA Goddard Space
Flight Center, with inputs from the Rio de Janeiro City Government in order to facilitate the

2
system customization for operations. Dr. Dalia Kirschbaum will act as the NASA/GSFC sponsor
and point of contact and Dr. Augusto Getirana will lead the science requirements of the project.
While it is expected that the final product of this project will provide valuable and actionable
information on floods in Rio de Janeiro, the efforts in developing such a modeling system will
serve as a pilot study for future application to other cities facing similar issues.

Workplan

The plan we present here has four components: 1) retrospective analysis using readily available
model parameters (2) retrospective analysis with customized parameters; 3) customization of a
2-D hydrodynamic model; and 4) transfer of modeling system to local partners for operations.

1) Retrospective analysis using readily available model parameters (6 months)

As the first step a pilot study is designed using readily available datasets for LIS. This means that
default model parameters will be used in order to define a benchmark for our further model
customization. For our pilot, the following tasks are required:
1. Configure LIS (LSMs and HyMAP) to run over Rio de Janeiro city using IMERG as input for
the period of record (2014-present) on a 0.5-km or 1-km grid. (2 month)
2. Compare the modeled soil moisture and evapotranspiration from IMERG runs with SMAP
and ALEXI data, and compute statistics based on changes in soil moisture and the ability
to detect events as well as ability to miss them. (2 months)
3. Compare modeled streamflow with available observations. (1 month)
4. Compile statistics on probability of detection and false alarm rates based on SMAP and
streamflow. (1 month)

2) Retrospective analysis with customized parameters (8 months)

Assuming the regional retrospective analysis using readily available model parameters is
successful, then the next step will be to customize as many parameters as possible using datasets
provided by the local partners. They are:
1. Process LiDAR data to be used as digital elevation model in LIS (2 month)
2. Process high resolution land cover map to be used in LIS (1 month)
3. Generate ground-based precipitation fields using interpolation techniques, such as
Krigging, and compare them against IMERG. The ground-based product can be further
merged with satellite-based data towards an improved precipitation product for the
region. (2 months)
4. Run LIS models over the same domain with customized datasets and repeat tasks 2-4 in
step 1). (3 months)

3) Customization of a 2-D hydrodynamic model (4 months)

Assuming the retrospective analysis with customized parameters performed better than the
benchmark, the following step is the customization of a 2-D hydrodynamic model. In
collaboration with Dr. Paul Bates, the LISFLOOD-FP model will be customized for the city of Rio

3
using in situ data provided by the Rio de Janeiro City Government. Simulations will be performed
for extreme events occurred in recent years and evaluated using ground-based observations.

4) transfer of modeling system to local partners for operations (2 months)

The final component is the transition of the system for operations to IPP. Final tests will be
perform locally ensuring that the system will work. We also plan to train staff and build capacity
at IPP, Fundação Rio Águas, Fundação GeoRio and Rio Operations Center. This training will focus
on the use for operations of the system.

Four additional months are reserved in this workplan for publication writing and eventual delays
in executing the tasks.

Workforce Requirements
We anticipate that one full-time person could complete the proposed activities within the
feasibility study the work could build off the availability and expertise of a small group. We
anticipate that once the feasibility study is completed and the real-time global analysis has been
configured, the requirements for maintaining this system (other than PPS support) will be
provided by the Brazilian partners.

References

Bates, P. D., M. S. Horritt, and T. J. Fewtrell (2010), A simple inertial formulation of the shallow
water equations for efficient two-dimensional flood inundation modeling, J. Hydrol.,
387, 33–45, doi:10.1016/j.jhydrol.2010.03.027.
Bates, P.D., De Roo, A.P.J., 2000. A simple raster-based model for flood inundation simulation.
Journal of Hydrology 236 (1–2), 54–77.
Getirana, A., A. Boone, D. Yamazaki, B. Decharme, F. Papa, and N. Mognard, 2012: The
Hydrological Modeling and Analysis Platform (HyMAP): Evaluation in the Amazon basin.
J. Hydrometeor., 13, 1641–1665, doi:10.1175/JHM-D-12-021.1.
Getirana, A., Peters-Lidard, C., Rodell, M., Bates, P.D., 2017. Trade-off between cost and
accuracy in large-scale surface water dynamic modeling. Water Resources Research.
DOI: 10.1002/2017WR020519
Kumar, S.V., C.D. Peters-Lidard, Y. Tian, P.R. Houser, J. Geiger, S. Olden, L. Lighty, J.L. Eastman,
B. Doty, P. Dirmeyer, J. Adams, K. Mitchell, E.F. Wood, and J. Sheffield, 2006: Land
Information System - An Interoperable Framework for High Resolution Land Surface
Modeling. Environmental Modelling & Software, 21, 1402-
1415, doi:10.1016/j.envsoft.2005.07.004
Kumar, S.V., C.D. Peters-Lidard, Y. Tian, P.R. Houser, J. Geiger, S. Olden, L. Lighty, J.L. Eastman,
B. Doty, P. Dirmeyer, J. Adams, K. Mitchell, E.F. Wood, and J. Sheffield, 2006: Land
Information System - An Interoperable Framework for High Resolution Land Surface
Modeling. Environmental Modelling & Software, 21, 1402-
1415, doi:10.1016/j.envsoft.2005.07.004

4
Ozdemir H., Sampson C.C., de Almeida G.A.M., Bates P.D. (2013). Evaluating scale and
roughness effects in urban flood modelling using terrestrial LIDAR data. Hydrology and
Earth System Sciences, 17, 4015-4030.
Sampson C.C., Fewtrell T.J., Duncan A., Shaad K., Horritt M.S. , Bates P.D. (2012). Use of
terrestrial laser scanning data to drive decimetric resolution urban inundation
models. Advances in Water Resources, 41, 1-17.