Wednesday, February 21, 2018

The GRASS module r.recovery

The r.recovery module was implemented as an add-on for GRASS software, a largely applied open-source Geographic Information System (GIS), and it offers a complete environment for the simulation of forestry regeneration in conservation areas, including a built-in tool for calibration and validation of the model parameters by means of standard and freely available satellite imagery. Besides, it allows the delineation of irregular computational domains, such that the user is allowed to precisely select arbitrary portions of the scene where the simulation should be performed. We remark
that the r.recovery tool can be advantageously applied by forestry managers and policy-makers as a form of acquiring technical and scientifically-based information for strategy development and decision-making.

Program name: r.recovery.
Developer: L.A. Richit.
Contact address: luizaugustorichit@gmail.com.
Year first available: 2018.
Software required: GRASS GIS 7.0 or later.
Program language: C.
Package size: 144 kB (source code).

Download

Monday, October 9, 2017

GPGPU-accelerated solution implemented using CUDA C/C++

We are proud to share here our more recent work on environmental modelling. We present a GPGPU implementation of finite-differences solution of the equations of the 2D groundwater flow in unconfined aquifers for heterogeneous and anisotropic media.

Here the code information:

Program name: ParGW
Developer: Tomas Carlotto.
Contact address: thomas.carl@hotmail.com
Year first available: 2017
Software required: CUDATM Toolkit 8.0 or later and CUSP library.
Program language: CUDATM C++.
Cost: free of charge.


Download

Bellow two animations showing the dynamic of soil saturation in a basin. Both were modeled with our implementation.


 

Monday, March 14, 2016

Flow of water in a soil profile: a gravity dam simulation

We present here a task carried out by our students in the Environmental Modeling subject. They implemented a routine in Matlab/Scilab/Octave high-level interpreted language in order to simulate the flow of water in a soil profile. Their hypothetical study case was a gravity dam. They performed two simulations 1) steady simulation, with the reservoir water level kept constant and 2) unsteady simulation, with the level changing from 10 meters to 50 meters and form 50 meters to 10 meters, whereas the downstream level was kept constant. See the pictures attached and download the pdf report.





Saturday, February 13, 2016

Artificial Neural Network Ensemble (ANNE) algorithm

As our first post here, we would like to share our Artificial Neural Network Ensemble (ANNE) algorithm implemented in Scilab. We have used it as part of our studies on riparian buffer design (see picture bellow). Feel free to download it and let us know your comments, critics and suggestions.

Also download here the paper:
Santin FM, Grzybowski JMV, Silva RV. 2015. Application of neural network ensembles to the problem of estimating raparian buffer width as a function of desired filtering properties. In Ist International Congress of Management, Technology and Inovation. Erechim; 1–4.