Artificial neural networks and multiple linear regression model using principal components to estimate rainfall over South America

Soares dos Santos, T.; Mendes, D.; Rodrigues Torres, R.

doi:https://doi.org/10.5194/npg-23-13-2016

Articles | Volume 23, issue 1

https://doi.org/10.5194/npg-23-13-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

https://doi.org/10.5194/npg-23-13-2016

© Author(s) 2016. This work is distributed under
the Creative Commons Attribution 3.0 License.

Articles | Volume 23, issue 1

Research article

|

27 Jan 2016

Research article |

| 27 Jan 2016

Artificial neural networks and multiple linear regression model using principal components to estimate rainfall over South America

T. Soares dos Santos, D. Mendes, and R. Rodrigues Torres

Download

Final revised paper (published on 27 Jan 2016)
Preprint (discussion started on 06 Aug 2015)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC C434: 'First comments', Anonymous Referee #1, 21 Sep 2015
- AC C574: 'Authors' reply', Thalyta Santos, 08 Dec 2015
RC C450: 'Anonymous #2 referee comments', Anonymous Referee #2, 25 Sep 2015
- AC C576: 'Authors' reply', Thalyta Santos, 08 Dec 2015

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Thalyta Santos on behalf of the Authors (01 Jan 2016) Author's response Manuscript

ED: Referee Nomination & Report Request started (07 Jan 2016) by Vicente Perez-Munuzuri

RR by Anonymous Referee #1 (07 Jan 2016)

ED: Publish as is (07 Jan 2016) by Vicente Perez-Munuzuri

AR by Thalyta Santos on behalf of the Authors (12 Jan 2016)

Short summary

Statistical downscaling is widely used in large operational centers around the world, using exclusively linear relations (MLR); this study uses a statistical downscaling methodology using a nonlinear technique known as ANNs with CMIP5 project data. The artificial neural network can perform tasks that a linear program cannot. The main advantages of this are its temporal processing ability and its ability to incorporate several preceding predictor values as input without any additional effort.