Research proposal
Introduction
Both climate and landcover have strong effects on river flow. Earth’s climate is warming and
landcover is changing via human interventions. As a consequence, river flow may change. However,
to date, causes of river flow changes remain strongly debated, and effects of landcover and climate
change remain difficult to predict. In this project I will quantify how river flow has changed across the
contiguous United States of America over the period since 1980, and will try to attribute these changes
to changes in either temperature, precipitation or landcover. When we know the main cause of stream
flow changes, we can predict how it will change in the future and if human intervention is needed
and/or possible.
Problem statement
Stream flow has a direct impact on the surrounding ecosystem, as changes in it can impact vegetation,
groundwater, nutrient supply and general water supply (Creed et al., 2014; Patil et al., 2022). In order
to correctly assess future water resources and how to manage them, we have to understand how
climate and landcover change can impact them (Berghuijs & Woods, 2016). This has been done in
several different researches, such as by Zhou et al. (2015) and Roderick and Farquhar (2011), but there
is a debate about how effective and precise these researches are. For example, according to Berghuijs
and Woods (2016), Zhou et al. make several assumptions about the temporal and spatial conditions
that make their framework inaccurate.
On the subject of landcover change specifically there has been made a base for further research by
Sterling et al. in 2012. They published a paper about the effect of anthropogenic land cover change on
annual runoff. Their research has found that of all drivers of annual global runoff, this human
interference could very well be the main one. It would be interesting to see if this finding also rings
true when looking at a smaller scale, compared to the global research they have done.
The goal of this research is to look at streamflow on a different spatial scale, and on a bigger and more
recent timescale than in previous research as to incorporate more detail and provide a more precise
representation of the real life situation. To realize these goals about our understanding on streamflow
changes in the USA, the following research questions have to be addressed:
1. How has river flow across the contiguous United States changed since 1980?
2. To what extent can we use year-to-year variations of climate and river flow to determine which
factors have controlled the observed streamflow trends?
3. To what extent can we use long-term trends of climate, river flow, and landcover to determine
which factors have controlled the observed streamflow trends?
These three sub questions lead up to an answer to the main research question: Are changing river
flows across the US caused by climate or vegetation?
Objectives
The main objective of this research is to provide a new dataset by combining two separate data sets
that can be used to accurately calculate river flow change and the mechanisms behind it. In order to do
this, the study has the following sub-objectives:
1. To improve the researcher’s skill in programming with Python;
, 2. To review whether this type of data can give insight to general streamflow change, instead of data
focused on a research area of a different size;
3. To provide a set-up for a similar dataset in other parts of the world, such as in Europe;
4. To help provide a base for future research on stream flow predictions and possible human
intervention and water management.
The result of this study will be valuable to the scientific community as well as to water management
policy makers in understanding stream flow and how to process stream flow data.
Preliminary Literature Review
The effect that changes in climate or land cover have on river flow has been researched on both a
global and a local basis.
In the research of Evaristo and McDonnell (2019) for instance, they made a global analysis on the
change in water storage and loss due to deforestation and reforestation. This paper got a response from
the scientific world in the form of a counter paper by Kirchner et al. (2020), calling the main results of
Evaristo unsubstantiated. The paper by Hornbeck et al. (1970) is an example of a smaller scale look at
the effect of deforestation on stream flow. They researched the change in river flow in New England in
a two year time period.
Another research in the United States is the paper by Hopkins et al. (2015) in which they looked at the
effect of urbanization on river flow.
In 2011, Wang et al. have looked at 413 watersheds in the contiguous United States of America to
quantify the effect of climate change and human impact on streamflow. The dataset they used, Model
Parameter Estimation Experiment (MOPEX), contains data from 1948 to 2003. This dataset is
extensive, but not so recent and thus not as suitable for predicting future stream flow in the USA.
Vicente‐Serrano et al. have looked at the same type of data in several European countries in the period
1961 to 2012.
Previous research based on the CAMELS dataset includes one from Huang et al. (2021). In this
research, the goal was to estimate long term and seasonal baseflow in the period 1981 - 2014. The
main difference between their and this research is that they focused on baseflow. Baseflow is a part of
streamflow, it does not take the direct input of precipitation in account. Also, Huang et al. got their
land cover data from the CAMELS dataset itself, whereas for this research two datasets will be
combined.
Another research based on the CAMELS data set is by Wu et al. that was published in 2020. The focus
of this study was to determine the dominant streamflow generation mechanisms at the event scale,
such as snowmelt and storm events.
So far, changes in river flow in the contiguous USA have not been securely attributed to a single
characteristic. There is an open space in the collective knowledge on how stream flow has changed in
the last decades that this research will try to fill.
Methodology
To start the research, the two datasets containing the hydrological, climatological and land cover data
have to be combined. For the hydrological and climatological data, the CAMELS data set will be used.
This can be downloaded from the National Center for Atmospheric Research (NCAR) website. The
land cover data will be derived from the Rangeland Analysis Platform by Google.
Combining these two datasets is done by loading the CAMELS shapefiles into QGIS. This can be
done manually, but will preferably be automatized by writing a code in Python and using Google
Earth Engine.
Introduction
Both climate and landcover have strong effects on river flow. Earth’s climate is warming and
landcover is changing via human interventions. As a consequence, river flow may change. However,
to date, causes of river flow changes remain strongly debated, and effects of landcover and climate
change remain difficult to predict. In this project I will quantify how river flow has changed across the
contiguous United States of America over the period since 1980, and will try to attribute these changes
to changes in either temperature, precipitation or landcover. When we know the main cause of stream
flow changes, we can predict how it will change in the future and if human intervention is needed
and/or possible.
Problem statement
Stream flow has a direct impact on the surrounding ecosystem, as changes in it can impact vegetation,
groundwater, nutrient supply and general water supply (Creed et al., 2014; Patil et al., 2022). In order
to correctly assess future water resources and how to manage them, we have to understand how
climate and landcover change can impact them (Berghuijs & Woods, 2016). This has been done in
several different researches, such as by Zhou et al. (2015) and Roderick and Farquhar (2011), but there
is a debate about how effective and precise these researches are. For example, according to Berghuijs
and Woods (2016), Zhou et al. make several assumptions about the temporal and spatial conditions
that make their framework inaccurate.
On the subject of landcover change specifically there has been made a base for further research by
Sterling et al. in 2012. They published a paper about the effect of anthropogenic land cover change on
annual runoff. Their research has found that of all drivers of annual global runoff, this human
interference could very well be the main one. It would be interesting to see if this finding also rings
true when looking at a smaller scale, compared to the global research they have done.
The goal of this research is to look at streamflow on a different spatial scale, and on a bigger and more
recent timescale than in previous research as to incorporate more detail and provide a more precise
representation of the real life situation. To realize these goals about our understanding on streamflow
changes in the USA, the following research questions have to be addressed:
1. How has river flow across the contiguous United States changed since 1980?
2. To what extent can we use year-to-year variations of climate and river flow to determine which
factors have controlled the observed streamflow trends?
3. To what extent can we use long-term trends of climate, river flow, and landcover to determine
which factors have controlled the observed streamflow trends?
These three sub questions lead up to an answer to the main research question: Are changing river
flows across the US caused by climate or vegetation?
Objectives
The main objective of this research is to provide a new dataset by combining two separate data sets
that can be used to accurately calculate river flow change and the mechanisms behind it. In order to do
this, the study has the following sub-objectives:
1. To improve the researcher’s skill in programming with Python;
, 2. To review whether this type of data can give insight to general streamflow change, instead of data
focused on a research area of a different size;
3. To provide a set-up for a similar dataset in other parts of the world, such as in Europe;
4. To help provide a base for future research on stream flow predictions and possible human
intervention and water management.
The result of this study will be valuable to the scientific community as well as to water management
policy makers in understanding stream flow and how to process stream flow data.
Preliminary Literature Review
The effect that changes in climate or land cover have on river flow has been researched on both a
global and a local basis.
In the research of Evaristo and McDonnell (2019) for instance, they made a global analysis on the
change in water storage and loss due to deforestation and reforestation. This paper got a response from
the scientific world in the form of a counter paper by Kirchner et al. (2020), calling the main results of
Evaristo unsubstantiated. The paper by Hornbeck et al. (1970) is an example of a smaller scale look at
the effect of deforestation on stream flow. They researched the change in river flow in New England in
a two year time period.
Another research in the United States is the paper by Hopkins et al. (2015) in which they looked at the
effect of urbanization on river flow.
In 2011, Wang et al. have looked at 413 watersheds in the contiguous United States of America to
quantify the effect of climate change and human impact on streamflow. The dataset they used, Model
Parameter Estimation Experiment (MOPEX), contains data from 1948 to 2003. This dataset is
extensive, but not so recent and thus not as suitable for predicting future stream flow in the USA.
Vicente‐Serrano et al. have looked at the same type of data in several European countries in the period
1961 to 2012.
Previous research based on the CAMELS dataset includes one from Huang et al. (2021). In this
research, the goal was to estimate long term and seasonal baseflow in the period 1981 - 2014. The
main difference between their and this research is that they focused on baseflow. Baseflow is a part of
streamflow, it does not take the direct input of precipitation in account. Also, Huang et al. got their
land cover data from the CAMELS dataset itself, whereas for this research two datasets will be
combined.
Another research based on the CAMELS data set is by Wu et al. that was published in 2020. The focus
of this study was to determine the dominant streamflow generation mechanisms at the event scale,
such as snowmelt and storm events.
So far, changes in river flow in the contiguous USA have not been securely attributed to a single
characteristic. There is an open space in the collective knowledge on how stream flow has changed in
the last decades that this research will try to fill.
Methodology
To start the research, the two datasets containing the hydrological, climatological and land cover data
have to be combined. For the hydrological and climatological data, the CAMELS data set will be used.
This can be downloaded from the National Center for Atmospheric Research (NCAR) website. The
land cover data will be derived from the Rangeland Analysis Platform by Google.
Combining these two datasets is done by loading the CAMELS shapefiles into QGIS. This can be
done manually, but will preferably be automatized by writing a code in Python and using Google
Earth Engine.
