Computer Resources



INPE has a supercomputing system installed at CPTEC headquarters in Cachoeira Paulista / SP called Tupã. It is a Cray XE6 equipment that was purchased with funds from the Ministry of Science, Technology, Innovation and Communications (MCTIC) and the São Paulo State Research Support Foundation (FAPESP) for shared use of CPTEC and CCST, both INPE, as well as research groups, institutions and universities that are part of MCTIC’s Brazilian Climate Change Research Network (CLIMA Network), the FAPESP Global Climate Change Research Program (PFPMCG) and the National Institute of Science and Technology for Climate Change (INCT for Climate Change).

With a top speed of 258 teraflops, equivalent to 258 trillion calculations per second, Tupã, at the time of its installation, was the 29th fastest supercomputer in the world and the 3rd most powerful supercomputer dedicated to operational numerical weather and weather forecasting. seasonal climate. Considering the applications for climate change, it ranked 8th. This ranking has placed Brazil among the countries capable of generating future climate scenarios in support of the IPCC’s fifth report, the UN panel that assesses the consequences of global warming. The new machine is also fundamental to the development and implementation of the Brazilian Model of the Global Climate System, which will incorporate all the elements. Further technical details of this system can be found at http://supercomputacao.inpe.br/documentacao2.

The CST Graduate Program has server machines and computers dedicated to the specific activities and tasks of students and teachers. This computer park also includes high-performance laser printers, scanning systems, poster printing plotters, banners, scientific and promotional materials, and IT helpdesk. This equipment is installed in the CCST buildings on the São José dos Campos and Cachoeira Paulista campus and has direct access to the supercomputing system.

The ELAT Laboratory has a park of 10 servers with more than 200 processors and 2 24Tb and 52Tb arrays for processing and storing data from lightning monitoring networks, as well as for generating regional weather model products. LABREN Laboratory has a cluster of 128 processors and 24Tb storage capacity used in the development and improvement of solar radiation and wind potential models. The TerraME-Galileo Laboratory enables high-performance computing, based on computer cluster architecture, for research and development in integrated environmental change modeling and the analysis of its social and economic impacts. This lab interconnects with INPE’s other facilities in the area of ​​high-performance computing, supercomputing and modeling, in particular CPTEC and OBT and also with TerraLAB Laboratory at the Federal University of Ouro Preto (UFOP), associated with INPE for new technologies. Developments of the distributed modeling environment by expanding computational and methodological capabilities for the development, testing and experiments with increasingly complex models.

PPG’s CST student and faculty have access to the numerical modeling infrastructure and socio-environmental databases needed for Earth System modeling and environmental modeling. These are basically numerical systems and databases developed and / or used within the Center itself and other INPE coordinations such as CPTEC and OBT, as well as through scientific cooperation with several leading numerical modeling institutions in the world. In addition, the Center uses various databases currently available such as: meteorological and oceanic information, meteorological satellite images, satellite derived meteorological products, data from various meteorological, hydrological and environmental variables of the National Data Collection Platform Network (PCD). ) from INPE (http://bancodedados.cptec.inpe.br), data from lightning monitoring networks, social and environmental indicators. Among the computational tools necessary for its activities, PPG uses various software developed or licensed at INPE such as:

ENVI: Licensed software used for digital processing of remote sensing images.

SPRING: State-of-the-art Geographic Information System (GIS) with functions of image processing, spatial analysis, numerical terrain modeling and query to spatial databases (http://www.dpi.inpe.br/spring) )

TerraLIB: GIS class and function library, available on the Internet as open source, enabling a collaborative environment for the development of GIS applications and tools (http://www.dpi.inpe.br/terralib)

TerraView: Application built from the TerraLib geoprocessing library that handles vector (points, lines, and polygons) and matrix (grids and images) data, both stored in market relational or geo-relational database systems, including ACCESS, PostgreSQL , MySQL and Oracle (http://www.dpi.inpe.br/terraview)

TerraME: Free and open source tool (http://www.terrame.org) for the development of dynamic spatial models developed by INPE in partnership with the Federal University of Ouro Preto (UFOP). This tool aims to support the construction of models that simulate the interactions of human and environmental systems. It provides a number of generic concepts that can be used for modeling in different areas of knowledge, and enables the use of real data in simulations through integration with spatial databases using DPI / INPE-developed TerraLib technology. . It is a dynamic modeling environment based on cellular automata that links cellular spaces to geographic databases for storage and retrieval. In this environment, land use change models, hydrological models, vegetation models, biodiversity models, and models of the various components of the terrestrial system can be developed.

Integrated Land Surface Model (INLAND): Integrates relevant surface and natural and agricultural ecosystem processes, which represent not only surface effects on climate, but also the effect of climate on surface processes (ecosystem dynamics, agricultural productivity). ). The model can be forced with site-specific weather data or grid-point climate data, or coupled with regional or global atmospheric circulation models. INLAND is developed with the participation of Universities and National Research Centers, State Research Networks and international collaboration, and is under the coordination of the National Institute for Space Research (INPE) and Federal University of Viçosa (UFV).

LuccME Model: Consists of a spatially explicit land use modeling framework developed by CCST and collaborators as an extension of the TerraME modeling environment. With LuccME you can easily build models of deforestation, agricultural expansion, desertification, forest degradation, urban growth, and other land use change and land cover processes at different scales and areas of study by combining available components or creating new ones, if deemed necessary.

INPE-EM Model: Developed by CCST itself in partnership with several institutions using TerraME modeling environment, it combines spatially explicit biomass and deforestation maps for a study area. The main feature of the INPE-EM framework is the flexibility in the emission process representation, allowing the refinement of the estimates according to the available data. The data are organized into regular cells whose resolution depends on the biomass and deforestation maps available in a given region. Both 1st order estimates (assuming 100% of emissions occur at the time of the use / coverage transition) and 2nd order estimates (considering the gradual process of carbon release and absorption, representing flows between compartments) can be generated. biomass). All parameters representing emission / absorption processes (for 1st and 2nd Order estimates) can be spatially explicit, making it possible to represent spatial heterogeneity within a region (eg, emission factors). Finally, the framework is easily parameterized, extensible and open source, allowing the development of new modules or modification of existing ones.

BRASIL-SR Model: Physical model for estimating surface incident solar radiation, which combines the approximation of “two-streams” in the solution of the radiative transfer equation with the use of statistically determined parameters from satellite images. . Cloud cover is considered the main factor of modulation of atmospheric transmittance, and the other optical properties are parameterized based on meteorological variables of surface temperature, relative humidity, atmospheric visibility and surface albedo. Estimates of incident solar irradiation on the surface (in Wh / m2) are made for the horizontal global, direct normal, horizontal diffuse and global components in the inclined plane. The model validation results show that it provides reliable data on solar resource availability and variability, which are essential to support the development of projects for harnessing it as a source of energy.

Distributed Hydrological Model (MHD-INPE): Works as a parametric conceptual / semi-conceptual model that seeks to understand the patterns of a watershed in order to simulate water storage and runoff processes. The runoff generation methodology combines a statistical formulation of the soil storage distribution (Xinanjiang Model) with the TopModel philosophy to represent the effect of topography on the spatial organization of soil water content in the basin. Future perspectives are: application of new climate scenarios, incorporation of LUCC-ME land use scenarios, hydrology impact and vulnerability assessments.

The regional climate modeling group works in a multidisciplinary manner and aims to understand the natural variability of past climate as well as to conduct past, present and climate climate simulations with a focus on providing technical and scientific information on possible climate change that may occur. happen in the next century in South America with a view to climate change mitigation and adaptation strategies. The following models are used for this:

Global Atmospheric General Circulation Model (AGCM): A numerical model for simulating troposphere dynamics on a global scale, used operationally in CPTEC / INPE and essential for providing boundary conditions for other limited area models;

Brazilian Terrestrial System Model (BESM): version of the CPTEC global model coupled with the ocean model MOM3 and which is under development;

ETA Model: It has an institutional history, being the meteorological model used for CPTEC weather forecast since the 90’s. It has technical support within INPE, its source code is under constant development by CPTEC. The climate change model version was physically developed for South America and is used to generate the scenarios that made up the National Communications on Climate Change. It also has a paleoclimatic version available. It is a computationally efficient community model with public domain documentation and is currently used in several Brazilian research institutions: UFLA, UFRJ, UEA, INPA, FUNCEME, UNB, COPPE, UENF, UPF. Its main technical features are: (i) domain covers most of South America and can be run in hydrostatic and non hydrostatic mode; (ii) Eta coordinate, suitable for mountainous areas such as South America; (ii) grid point model using grid E; (iv) prognostic variables: temperature, specific humidity, wind components u and v, surface pressure, turbulent kinetic energy and cloud water; (v) finite volume model (for long simulations); (vi) two parameterization schemes Betts-Miller-Janjic and Kain-Fritsch; (vii) Zhao and Ferrier microphysics schemes; (viii) Mellor Yamada 2.5 turbulence models; (ix) NOAH and NOAH-MP surface schemes; (x) computational dynamics with two levels of integration in time, which avoids spurious solutions.

RegCM4 Model: Limited area regional model developed by ICTP of Italy, with source code constantly improving and updating. Community, has easy installation and operation, very good computational performance and good documentation. It has an open discussion and support forum on the Internet and is available from the IPCC Global Template Database. It can be easily adapted to enable climate studies, paleoclimatic studies and climate projections. It is currently used by several Brazilian institutions: USP, UFPA, UFV, INIFEI, UECE, UNESP. Its technical characteristics are: (i) integration with RegESM framework; (ii) post processing in NetCDF format; (iii) MPI2 library; (iv) non-hydrostatic MM5 dynamic core; (v) vertical coordinate Sigma-p; (vi) Arakawa-B horizontal grid; (vii) time integration with split-explicit scheme; (vii) topography GMTED dataset; (viii) chemistry model DUST 12 bin; (ix) BATS, CLM 3.5 and CLM 4.5 surface models; (x) global data repository available at http://clima-dods.ictp.it/regcm4; (xi) schemas of fixed, relational lateral boundary conditions (linear and exponential techniques), time dependent, time dependent and input and output flow, sponge.

CCATT-BRAMS Model (“Chemistry Coupled Aerosol and Tracer Transport”): Real-time operational model of transport system monitoring using a 3D model coupled to an emission model.

GPSAS / RPSAS: Data assimilation codes developed by the Data Assimilation Office (DAO), currently GMAO Global Modeling and Assimilation Office (GMAO) of NASA, USA.

Potential Vegetation Model (PVM): Used to simulate the potential distribution of biomes from climate scenarios.

WWATCH III (“Ocean Wave Model”): Third generation ocean wave model developed by NOAA / NCEP (USA). This model is extensively used for regional and global scale forecasting in various weather centers around the world (ECMWF, NMC, etc.).