Abstract
This work represents an amalgam of a group of studies with the purpose of understanding the influence of the Pacific and Atlantic sea surface temperature on precipitation and river discharge in Northeastern South America.
Sea surface temperature is a good representative of phenomena such as ENSO that, in turn, cause worldwide climate variability. The patterns of the Atlantic Ocean sea surface temperature anomaly also plays a very important role in the precipitation over the neighboring regions and a special investigation was also carried out to better understanding this influence. The influence of these oceans' sea surface temperature on the intraseasonal variability of precipitation in Northeast Brazil was also a particular subject of study.
Statistical methods were largely used both during these investigations and in the development of models for forecasting discharge long term in advance at some sites in the Amazon, Orinoco and Tocantins River Basins.
Sea surface temperature anomalies in both oceans significantly influence precipitation over northeastern South America. The Atlantic Ocean, however, plays a more important role in the case of precipitation over Northeast Brazil while the Pacific Ocean seems to have stronger influence over eastern and northern Amazonia. As a result of changes in precipitation, the river discharge in the Amazon Region is also influenced by changes in sea surface temperature patterns.
The discharge of rivers located to the north of the Amazon River is mainly influenced by the Pacific sea surface temperature while the Atlantic influences the rivers to the south of the Amazon River. This influence could be clearly observed using the forecast models.
Two different methodologies were used to develop forecast models: Canonical Correlation Analysis and Artificial Neural Network. The first is a linear technique and the second a non-linear one. In both cases, the models developed using Pacific sea surface temperature were better at forecasting discharge at sites to the north of the Amazon River and those developed from Atlantic sea surface temperature at forecasting discharges at sites to the south of the Amazon River.
Even though the use of a non-linear technique improved the accuracy of the models in general, it considerably improved the capacity of Atlantic sea surface temperature to forecast discharge. This general improvement was to some degree expected considering the very complex and non-linear mechanisms that transform precipitation into discharge.
Sea surface temperature is a good representative of phenomena such as ENSO that, in turn, cause worldwide climate variability. The patterns of the Atlantic Ocean sea surface temperature anomaly also plays a very important role in the precipitation over the neighboring regions and a special investigation was also carried out to better understanding this influence. The influence of these oceans' sea surface temperature on the intraseasonal variability of precipitation in Northeast Brazil was also a particular subject of study.
Statistical methods were largely used both during these investigations and in the development of models for forecasting discharge long term in advance at some sites in the Amazon, Orinoco and Tocantins River Basins.
Sea surface temperature anomalies in both oceans significantly influence precipitation over northeastern South America. The Atlantic Ocean, however, plays a more important role in the case of precipitation over Northeast Brazil while the Pacific Ocean seems to have stronger influence over eastern and northern Amazonia. As a result of changes in precipitation, the river discharge in the Amazon Region is also influenced by changes in sea surface temperature patterns.
The discharge of rivers located to the north of the Amazon River is mainly influenced by the Pacific sea surface temperature while the Atlantic influences the rivers to the south of the Amazon River. This influence could be clearly observed using the forecast models.
Two different methodologies were used to develop forecast models: Canonical Correlation Analysis and Artificial Neural Network. The first is a linear technique and the second a non-linear one. In both cases, the models developed using Pacific sea surface temperature were better at forecasting discharge at sites to the north of the Amazon River and those developed from Atlantic sea surface temperature at forecasting discharges at sites to the south of the Amazon River.
Even though the use of a non-linear technique improved the accuracy of the models in general, it considerably improved the capacity of Atlantic sea surface temperature to forecast discharge. This general improvement was to some degree expected considering the very complex and non-linear mechanisms that transform precipitation into discharge.
Original language | English |
---|---|
Qualification | Doctor |
Awarding Institution |
|
Supervisors/Advisors |
|
Award date | 1998 Feb 16 |
Publisher | |
Publication status | Published - 1998 |
Bibliographical note
Defence detailsDate: 1998-02-16
Time: 10:15
Place: Room V:B
External reviewer(s)
Name: Hastenrath, Stefan
Title: Dr
Affiliation: Dept Atm. & Oceanic Sciences, University of Wisconsin-Madison
---
Subject classification (UKÄ)
- Water Engineering
Free keywords
- kartografi
- marklära
- geomorfologi
- Fysisk geografi
- climatology
- pedology
- cartography
- Physical geography
- geomorphology
- sea surface temperature
- rainfall-runoff
- Amazonia
- precipitation
- Northeast Brazil
- klimatologi
- Geophysics
- physical oceanography
- meteorology
- Geofysik
- fysisk oceanografi
- meteorologi
- Hydrogeology
- geographical and geological engineering
- Hydrogeologi
- teknisk geologi
- teknisk geografi