0
University of Amsterdam
Thesis: 30EC
THE RELATIONSHIP BETWEEN
ROSSBY WAVE ACTIVITY AND
CLIMATE CHANGE
Milan Verploegen (10001696) - MSc Earth Sciences (EM)
Supervisor: J.H. van Boxel
Co-assessor: B. Jansen
2017
1
Abstract
Rossby waves with larger amplitudes can create blocking flow patterns, resulting in stationary high-pressure systems. Stationary high-high-pressure systems lead to a high count of consecutive days with similar weather, increasing the frequency of events such as droughts, heat waves, cold spells or floods. This research aims to clarify the relationship between Rossby Wave Activity (RWA) at 500 hPa and climate change, to better understand the underlaying mechanics to changing climatological trends. Matlab is used to analyze reanalysis data provided by the ECMWF. The variables that are used in the analysis are: temperature, precipitation, the U- and V-component of wind speed at 500 hPa and the geopotential height at 500 hPa. A new RWA definition is created, that considers the wind speed and the direction of the wind relative to a purely westerly wind. This definition has proven to be valuable in looking for correlations with other variables. RWA shows an insignificant negative trend in time, but is strongly correlated to geopotential height and the difference in geopotential height between the sub-tropics and sub-Arctic regions, but less so to precipitation and temperature. The significance in correlations is influenced by the seasons; highest in winter, lowest in summer. Trends in temperature, precipitation and the difference in geopotential height are increasing between the sub-Arctic and sub-tropic regions on the Northern Hemisphere. The relationship between RWA and climatological trends is present but since there is no statistically significant trend in RWA, it is yet to be confirmed that this correlation is influencing trends in climate change.
2
Table of Contents
Abstract ... 1
1. Introduction ... 3
2. Research Questions ... 5
Main Research Question ... 5
Sub Research Questions... 5
3. Theoretical Framework ... 5
4. Methods ... 8
Step 1 – Data Collection ... 8
Step 2 – Data Processing ... 9
Step 3 – Data Analysis ... 13
5. Results ... 13 Trends ... 14 Correlations ... 17 Annual Correlations ... 17 Seasonal RWA ... 18 Seasonal GHD ... 19 6. Discussion ... 21
Rossby Wave Activity ... 21
Trends ... 21 Correlations ... 23 Reanalysis Data ... 24 Recommendations ... 24 7. Conclusion ... 25 References ... 26
Appendix I: Data Collection ... 31
Appendix II: RWA & GHD ... 33
Appendix III: Trends ... 34
3
1. Introduction
Climate change is known to cause rising sea levels, increased temperatures and species extinction (Cox et al., 2000 & Pereira et al., 2012). Over the past decades the issue has continuously gained more awareness among citizens, companies, organizations and political leaders (Lee et al, 2015). Businesses that do not follow the trend of going green are falling behind (Martínez-del-Rio et al., 2014). Taxes on plastic bags, increased prices on a full tank of gas and other signs of change are examples of climate change awareness making its way into society.
Every 5-6 years the IPCC (Intergovernmental Panel on Climate change) brings together all new research on climate change. According to the IPCC (2013b), global temperatures from 1951 to 2012 show a rise of 0.12°C per decade. Almost everywhere on the planet temperature is rising at an increasing rate. The rise in temperature is stronger on land than above the oceans (Seneviratne et al., 2016), meaning that an estimate of 0.12 degrees of decadal increase from 1951 to 2012 is most likely to be an underestimation when looking at climate change in urban and rural areas. A global decadal temperature increase of 0.12 degrees could mean a decadal temperature increase of 0.2°C, 0.3°C or in certain areas even over 0.4°C (IPCC, 2013b).
The changing climate results in more than just economic losses. Longer lasting and more intensive heat waves and cold spells are a threat to communities. The heat wave in Europe in the summer of 2003 led to a death toll of over 70.000 (Robine et al., 2008). The most occurring natural hazards, however, are floods, which occur more frequently with global warming as well as with the growth of the human population. Between 1992 and 2001 there have been close to a thousand reported floods worldwide, reaching a total death toll of nearly 100.000 and affecting the lives of over a billion people (McMichael et al., 2006). Diseases also occur more frequently as climate change creates better environmental settings for infectious diseases such as AIDS, Ebola, Lyme disease, toxic Escherichia coli and others (Epstein, 2001).
When looking at our knowledge about how the world is connected, there is still much to be done (Bony et al., 2006). One of the topics that can help us understand how our climate works, is atmospheric Rossby waves, planetary meandering waves that occur in the upper troposphere at pressure levels of approximately 500 to 300 hPa, roughly 5 to 10 km altitude (Aguado & Burt, 1999). The concept of Rossby waves at 500 hPa is not recent, publications date back to 1939 (Rossby, 1939). A visualization of the phenomena can be seen in figure 1. The colors indicate wind speed at 500 hPa, with red being higher speeds, followed by yellow, green and blue as decreasing and lowest speeds.
Wang et al. (2013) explain that Rossby waves play a key role in initiating low pressure systems that in turn influence daily weather conditions. High pressure systems caused by Rossby waves could be linked to some of the adverse effects of climate change such as heat waves or prolonged periods of increased precipitation. The driving forces behind Rossby waves are the combination three factors. The pressure difference between the tropics and the Arctic region creating flow between North and South, the Coriolis effect giving these flows a West-East component and vorticity causing the rotation of air which transforms the airflow into a meandering wave (Wang et al., 2013).
4
Figure 1 Visualization of Rossby Wave Activity over the Northern Atlantic and Europe.
Screenshot of August 10
th, 2016 at 11AM (Ventusky, 2017).
A better understanding of the relationship between Rossby Wave Activity (RWA) and climate change improves the insights in how our climate works. This insight in turn helps policymakers in dealing with negative consequences of climate changes, such as the Elbe flood in 2002 (Ulbrich et al., 2003 & Oetken et al., 2005) and the European heatwave in 2003 (Kreibich et al., 2005 & Stott et al., 2004). More knowledge on climate change will make it easier to perform risk assessments in for example
construction and urban planning, where changes in climate can decrease the durability of materials used in construction (Wang & Wang, 2012). Agriculture might require drastic changes in the type of crops. An increase in temperature and therefore heat waves (Russo et al., 2014) or frequent periods of drought for example can have severe consequences on crops that have little resistance to elevated temperatures and longer periods of water shortage (Anita et al., 2010 & Olesen et al., 2011).
It is poorly understood how the changing pressure difference caused by climate change affects patterns in RWA. Research on climate change and research on RWA is widely available (Becker et al., 2004, Peters & Vargin, 2015, Lillo & Parsons, 2017 & Zhao et al., 2017), research on the relation between the two topics however is scarce. With Arctic temperatures rising faster than mid-latitude and tropical
temperatures (Francis & Vavrus, 2012 & Biastoch et al., 2011), the slope of the 500 hPa level between the Arctic regions and the tropics is expected to be decreasing, thus slowing down mid altitude wind speeds in the temperate latitudes. The relation between Rossby waves and pressure systems at ground level are likely to be the cause of prolonged weather extremes (Petoukhov et al., 2013). This novel research seeks to provide additional knowledge to help gain insights to that relationship.
The aim of this research is to establish the influence of climate change on Rossby Wave Activity in the Northern Hemisphere and consequently the effect thereof on climatological trends. The research questions to complete this aim are specified in detail below.
5
2. Research Questions
Main Research Question
What is the relationship between Rossby waves at the 500 hPa level and climatological trends at temperate latitudes on the Northern Hemisphere?
Sub Research Questions
• How can RWA be defined?
• What are the trends in RWA, temperature, precipitation, zonal- and meridional wind speeds, geopotential height and the difference in geopotential height of the 500 hPa plain between the sub-tropics and the sub-Arctic regions in the Northern Hemisphere?
• How does RWA correlate to temperature, precipitation, geopotential height and geopotential height difference?
• How does the geopotential height difference correlate to temperature and precipitation?
3. Theoretical Framework
In this section, theoretical background information is provided on the building blocks of Rossby waves. Core concepts that are briefly explained are geopotential height and vorticity.
At the same pressure warm air will have a lower density than cold air. When air is lighter than the surrounding air, it will rise. At higher altitude levels pressure decreases, so the air will expand. This adiabatic (without the exchange of energy) expansion will cause the air to cool (Aguado & Burt, 1999). The tropics are warmer and contain air with lower density than regions in higher latitudes. This leads to airflows between the tropics and the Arctic regions (Aguado & Burt, 1999). Figure 2 illustrates a
summary of this movement.
6
Due to the rotation of the earth, air moving from the tropics toward the Arctic regions is adjusted toward the east. This is called the Coriolis effect. Adding relative vorticity results in Rossby waves, meandering streams of air that occur at approximately 5 to 10 km high, at the 500 to 300 hPa plane (Wang et al., 2013). The wind speed in this stream can be divided in two components; a South-North component (meridional) and a West-East component (zonal). These are called the V-component of wind speed and the U-component of wind speed respectively. The main direction of wind flow in Rossby waves is from West to East. Due to the meandering nature of Rossby waves, constantly alternating between South-North and North-South, it is possible for the direction of smaller segments of the flow to change from Eastward to Westward.
Geopotential height (GH) is the term used to describe the altitude of a pressure level above sea level. This research focusses on the GH of the 500 hPa plain, roughly the same height as where the Jetstream starts to occur. On average the 500 hPa level is found at approximately 5.6 km elevation. In colder zones, the 500 hPa plain often occurs around 5.2 km, while at warmer zones such as the tropics the same pressure is found at heights of over 5.5 km. The data used in this thesis indicates values for the 500 hPa level that range between 4777 m altitude (minimum at 60-70°N) and 5846 m altitude (maximum at 30-40°N).
Absolute vorticity is the sum of relative vorticity and planetary vorticity. Relative vorticity is the rotation of an object, or an amount of air relative to the earth surface. Planetary vorticity is the spinning of an object due to the rotation of the earth. Because the earth rotates around its axis, objects placed on the earth also rotate. The horizontal component of this rotation is strongest at the poles and decreases toward the equator, where the planetary vorticity is zero.
A simple experiment can be performed to visualize vorticity. The experiment requires a chair that can rotate. When rotating while sitting in the chair, you rotate at a certain speed. Making yourself smaller while doing this results in a faster speed. Doing the opposite by stretching arms and legs, the chair will rotate slower. Retracting arms and legs once more results in the chair picking up speed again.
In the Jetstream, vorticity mechanics applied to air are slightly different. Because near the 500 hPa plane absolute vorticity is a conserved quantity, relative vorticity increases when planetary vorticity decreases. This results in an increased amount of relative vorticity closer to the equator and a lower amount of relative vorticity near the poles (Aguado & Burt, 1999).
Figure 3 illustrates rotation caused by vorticity. When air flows back and forth between latitudes due to differences in GH, the air is subjected to different quantities of planetary vorticity, causing the air to start rotating. Air moving from high to low latitudes is subjected to a counter-clockwise vorticity, while air moving from low to high latitudes is subjected to clockwise vorticity. This rotation results in Rossby waves (NWS, 2017).
7
Figure 3 Example of spin (vorticity) resulting from the rotation of the earth (NWS, 2017).
Rossby waves are constantly moving but a wave can also stay stable for a period of several days. When this happens, high- and low-pressure systems created by respective down- and upward movement of air resulting from vorticity can become stationary. This strongly affects weather.Climate change causes global temperatures to rise. This rise in temperature is not evenly spread (Seneviratne et al., 2016). Some areas experience a higher increase in temperature than others. IPCC (2013a) shows temperatures rising faster above land than above the oceans. Due to the melting of ice, the Arctic regions show a steeper rise in temperature than the already warm tropics (Vaughan et al., 2003). This is called polar amplification. When the temperature is rising faster in the Arctic than in the tropics, a decrease in geopotential height difference (GHD) is to be expected, which in turn slows down wind speeds between the two regions.
Stagnating wind movement results in Rossby Waves that move slower. Blocking high pressure systems increase in longevity as a result. This has a direct effect on events such as heat waves. Examples are heat waves occurring more frequently in for example China (Sun et al., 2017), cold winters in the United States (Wallace et al., 2014 & Van Oldenborgh et al., 2015) and droughts in central and southwest Asia (Barlow et al., 2002) or Syria (Kelley et al., 2015).
8
4. Methods
Step 1 – Data Collection
Data for this research is obtained through the European Centre for Medium-Range Weather Forecasts (ECMWF, 2017). The ECMWF has a web service that provides free access to a wide variety of data. The reanalysis data used in this research are part of a larger dataset, called 20C (ECMWF, 2017). CERA-20C includes reanalysis data on multiple topics: the atmosphere, ocean waves, ocean sub-surface variables, land surface variables and sea ice, as well as an observation feedback archive. The range of the dataset covers a period from 1901 to 2010. Due to the size of datasets, a timespan of only 50 years of this dataset has been used. This covers the period of 1960 to 2010.
Besides the timespan, several other options must be selected regarding the downloading of the required datasets. The precise steps to downloading the exact same datasets used for this research can be found in the ‘Appendix I: Data Collection’. A summary of the type of data downloaded and used is provided in table 1.
Table 1 Summary of settings selected when requesting data download from ECMWF (2017).
Variable Units Level Type Frequency Step Latitude
Temperature K Surface
(2 m) Analysis 3-Hourly 0 40-70°N
Precipitation m Surface Forecast 3-Hourly 3-24
(per 3) 40-70°N Eastward component of wind speed: U m/s Pressure (500 hPa) Ensemble mean 3-Hourly 3-24 (per 3) 40-70°N Northward component of wind speed: V m/s Pressure (500 hPa) Ensemble mean 3-Hourly 3-24 (per 3) 40-70°N
Geopotential Height dm Pressure
(500 hPa) Analysis Monthly
3-24 (per 3)
30-40°N & 60-70°N
The following additional settings are the same for all variables and therefore not included in table 1: Longitude (180°W to 180°E), Period (1960 to 2010) and Format (NetCDF).
Before processing the downloaded data, a software program (Panoply, 2017) is used to create a simple and quick overview of the data. This program is used to view NetCDF files and can be used to ensure the integrity of NetCDF data. The software helps in making quick and simple plots of NetCDF data at any timestamp, which allow for a quick overview to verify that each step in the processing of the data has provided the sought-after results. It is also used to check if the downloaded files are complete, which is not always true. The larger the request at CERA-20C, the higher the chance that it either cannot be completed or that the data will be corrupt. Narrowing down the request by cutting the request into smaller timeframes or areas helps to prevent these issues, but it is recommended to make sure the files are in order after downloading by plotting them in Panoply (2017) or similar software.
9
Step 2 – Data Processing
Some of the processing is done by selecting the specific variables via CERA-20C. However, after downloading the files from CERA-20C, they still must be processed further before the data can be analyzed.
The following procedures in data processing and analyzing are performed using the Mathworks software called Matlab (2012) and saved as a Matlab script. The script in its entirety contains 1600 lines and over 10.000 words and is therefore attached to this document separately. Running the script in its entirety can cause memory issues. It is therefore recommended to run the script step by step, per segment, instead of all at once. This can be done by using the hotkey CTRL Enter in Matlab (2012). Especially the first segments require heavy computing power. In those segments, the raw NetCDF files are processed. Since this only contains few lines, it is possible to run the lines one by one, simply by copying them from the script to the command window. Using 16G DDR4 RAM and an i5 6600k Skylake Processor, running a single line from the first segment takes several minutes. To prevent freezes, it is recommended to convert the individual NetCDF files to 3D matrices separately. Once the 3D matrices are extracted from the NetCDF files, their size will be reduced significantly and the following steps will require less
computing power.
Before extracting the 3D matrices from the NetCDF files, the ‘ncdisp’ function is used to display the information required to extract the data. The function shows the dimensions and their size. In the case of temperature, this would be 360x31x149024 as longitude by latitude by time. Using the ‘ncread’ function, the variables are extracted from the NetCDF files. This decreases the size of the working variable but also removes additional information such as the starting time, dimension, size and unit. If preferred, the temperature matrix can be converted to Celsius.
The temperature, precipitation and U- and V-component data now exist out of 3D matrices of 360 by 31 and include eight points of data per day for 50 years. The mean of these points is calculated and added to a new file, thus creating matrices of daily data. The precipitation is an exception. Since the eight points per day are additive, only the 8th point is selected and added to a new file. Once the daily
matrices are created, the RWA variable can be calculated. Due to the large quantity of the data, an RWA definition is required that can be used to calculate the activity for the designated area and timespan. The definition created for this purpose considers the wind speed and the direction of the wind relative to a purely westerly wind. The wind speed (W) is calculated as follows:
𝑊 = √𝑈2+ 𝑉2. [m/s] (1)
The direction relative to a purely westerly wind (D) is calculated by using the arctangent: 𝐷 = 0.5 − 𝐴𝑡𝑎𝑛 ( 𝑈
|𝑉|+0.001) 𝜋⁄ [-] (2)
V is incremented by 0.001 m/s to avoid division by zero. The direction is normalized on a scale of 0 to 1 by dividing the result of the arctangent by π and subtracting the result from 0.5. Now a purely northerly or southerly wind (U = 0 m/s) will yield D = 0.5, meaning the wind direction is at a right angle with a westerly wind. A westerly wind (V = 0 and U > 0) will yield D = 0 and an easterly wind (V = 0 and U < 0) will yield D = 1.
10 Daily values of RWA are calculated by multiplying W and D:
𝑅𝑊𝐴 = √𝑈2+ 𝑉2∗ (0.5 − 𝐴𝑡𝑎𝑛 ( 𝑈
|𝑉|+0.001) 𝜋⁄ ) [m/s] (3)
This value is expressed in meters per second and is always positive. It increases in strength when there is a strong North to South or South to North flow. The value decreases when value of the V-component becomes small. The value on the right side of the multiplication sign reaches its maximum when the U-component is negative, but when this occurs the wind speed is generally low. Therefore, a wind directed from East to West does not result in a maximum RWA.
Monthly matrices of all parameters are created by averaging over the days of the months (summing in case of precipitation). Because months do not have an equal amount of days, an additional excel file is created, containing the number of days in every month from 1960 to 2010. This file is used to select the correct amount of days to calculate the monthly mean or sum.
The newly created monthly files are used to first calculate the GHD. The GHD matrix is created by subtracting the GH values of the area 60-70°N from the values of the 30-40°N area.
𝐺𝐻𝐷 = 𝐺𝐻 (30 − 40°𝑁) − 𝐺𝐻 (60 − 70°𝑁) [m] (4) Once the GHD monthly file is created, the monthly files are used to create seasonal files: spring, summer, autumn and winter. The four seasons are combined to also create the annual matrices. The seasons are defined in table 2.
Table 2 Definition of seasons
Season Spring Summer Autumn Winter
Months
March June September December April July October January May August November February
To limit the output of the correlation analysis and reduce noise in the data, the daily, seasonal and annual matrices are averaged over 10x10 degree areas, resulting in three rows (40°N to 70°N in steps of 10°) of 36 zones (180°W to 180°E in steps of 10°). These zones are in turn combined to create the areas that are used in the trend and correlation analyses. The areas are defined in table 3.
Table 3 Definition of areas
Region Longitudes Zones
World 180°W - 180°E 1 – 36 Europe 0°E - 60°E 18 – 23
Asia 60°E - 140°E 24 – 32 Pacific 140°E - 130°W 1 – 4 & 33 – 36 Canada 130°W - 60°W 5 – 12 Atlantic 60°W - 0°W 13 – 17 The annual data is visualized in figures 4 to 7.
11
Figure 4 Annual data containing 10x10 degree fields of Temperature (in °C) and Precipitation (in
mm) at 50-60°N. The Y-axis shows the year, ranging from 1960 to 2010. The X-axis shows the
longitude in degrees, ranging from -180° to 180°. The legend indicates the value of the colors.
Figure 5 Annual data containing 10x10 degree fields of GH (in m) at 30-40°N and 60-70°N and
GHD (in m) at 30-70°N. For axis and legend information see figure 4.
12
Figure 6 Annual data containing 10x10 degree fields of the Zonal (U) (in m/s) and Meridional
(|V|) (in m/s) components of the wind at 500 hPa, at 50-60°N. For axis and legend information
see figure 4.
Figure 7 Annual data containing 10x10 degree fields RWA (in m/s) at 40-50°N, 50-60°N &
60-70°N. For axis and legend information see figure 4.
13
Step 3 – Data Analysis
The first step in analyzing the data is included in the data processing: creating the RWA unit and calculating the GHD. These steps are performed in Matlab (2012). ‘Appendix II: RWA & GHD’ contains the parts of the script where these variables are calculated.
The trends are calculated by performing a regression analysis. The most important values calculated are the slopes and their significance. Optionally one can also calculate the slope per century and three significance values, which are added to gain a better understanding of the trends. The significance values only show a number for a significance less than 5%, 1% or 0.1% respectively. For more details, ‘Appendix III: Trends’ contains the part of the script where the trends are calculated.
The result of the trend calculations is a list of statistical information about the areas: World, Europe, Asia, the Pacific, Canada and the Atlantic. The trends per century and their significance are summarized and provided in the Results – Trends section.
Before the correlations are calculated, the files are re-organized to be able to use the ‘corr’ function in Matlab (2012). The matrices are reduced in size from 3D to a 2D and transposed, leaving a matrix with time as first dimension and area as second. This enables the software to use the built-in function to correlate the different variables based on their location.
5. Results
The results are divided into trends and correlations. The first section shows the trends, and the second section the correlations. The results are based on data from 1960 to 2010.
The following abbreviations are used in the results: • Temp. : Temperature
• Precip. : Precipitation
• RWA : Rossby Wave Activity
• |V| : Absolute value of meridional component of wind speed at 500 hPa • U : Zonal component of wind speed at 500 hPa
• GHD : Geopotential height difference • GH : Geopotential height
14
Trends
The following two tables show trends (Table 4) and their significance (Table 5). The trends are calculated per 100 years (c) using data from 1960 to 2010 and focus on 50°N to 60°N. Since GHD is calculated by subtracting the GH at 60-70°N from the GH at 30-40°N, GHD is noted as 30-70°N.
Table 4 Trends per century of different variables at 50-60°N for several areas. The definition of
the areas and abbreviations used can be found in Tables 2 and 3. Red numbers indicate negative
trends. Green marks a significance of P < 1% and yellow marks a significance of 1% < P < 5%.
Trends of Variables in Different Continents at 50-60°N per Century
Variable Area Temp [°C/c] Prec [mm/c] RWA [(m/s)/c] |V| [(m/s)/c] U [(m/s)/c] GHD 30-70°N [m/c] GH 30-40°N [m/c] GH 60-70°N [m/c] World 2.7 70 -0.11 -0.24 0.03 6.8 33 27 Europe 2.7 76 0.00 -0.04 0.36 12 29 18 Asia 3.0 96 -0.11 -0.22 0.29 -2.3 29 32 Pacific 2.0 8.8 -0.07 -0.82 -1.31 -1.8 24 26 Canada 3.4 62 -0.16 -0.48 -0.52 -3.2 38 41 Atlantic 1.7 109 -0.20 0.58 1.72 39 47 7.6
Table 5 Significance in % of the trends shown in Table 4. For the definition of the areas and
abbreviations see Tables 2 and 3 (Green: P <1%. Yellow: 1% < P <5%).
Significance in % of Trends of Variables in Different Continents at 50-60°N
Variable Area Temp [%] Prec [%] RWA [%] |V| [%] U [%] GHD 30-70°N [%] GH 30-40°N [%] GH 60-70°N [%] World 0.0 0.1 43 15 97 58 0.0 0.5 Europe 0.0 9.8 100 90 66 52 0.0 26 Asia 0.0 0.2 54 34 73 89 0.0 2.2 Pacific 0.0 79 80 3.7 25 93 6.8 3.9 Canada 0.0 7.6 38 13 54 85 0.0 0.3 Atlantic 0.7 1.0 43 13 30 17 0.0 73
The data of these trends regarding the area defined as ‘world’ are plotted and shown in figure 8. Trendlines are included.
15 Fi gu re 8 An n u al d ata p o in ts fro m 1 9 6 1 to 2 0 1 0 o f all vari ab les used in tabl e 2 and 3 ar e vi sua liz ed. The p lo ts sh o w the ar ea d efin ed as ‘w o rld ’, at 5 0 -60 °N , e xc ept f o r G eo p o te n tial H eig h t which is lo cat ed at 3 0 -70 °N . Trend lin es are pl o tt ed t o v isu ali ze the t rend s sho wn in T ab le 4 .
16
Temperature shows a significant increase for all areas. The temperature in all areas combined rises by 2.7°C/century. Europe also shows a 2.7°C/century increase. In Asia and Canada, the temperature rises faster, whereas the warming is slower over the Pacific and Atlantic oceans. The trend is very significant for all areas (P<1%).
For the area called world Precipitation shows an increase of 70 mm/century. All areas show an increase in precipitation. However, only the trends for World, Asia and the Atlantic are significant (P=0.1%, P=0.2% and P=1.0%). Europe and Canada show an increase in precipitation similar to that of the entire area, but these trends are not statistically significant. Remarkable is the low increase in precipitation over the Pacific of 8.8 mm/century.
RWA decreases for all areas except Europe, which does not show any trend at all. The trend is small in the Pacific, while the Atlantic shows the highest trend. However, for none of the areas the trend in RWA is significant.
The meridional component of wind speed (V) shows negative trends in all areas except the Atlantic. The only significant trend however is the Pacific, with a P-value of 3.7% and a trend of -0.82 (m/s)/century. The entire area shows a negative trend of -0.24 (m/s)/century however the significance only shows a P-value of 15%. Like RWA, the absolute V-component of wind speed also shows a trend P-value that is close to zero in Europe.
The zonal component of wind speed (U) is almost constant for the world. Europe, Asia and the Atlantic show an increase in zonal wind speeds. The Pacific and Canada show a decrease. The P-values indicate that the trends are insignificant for all individual areas. The lowest P-value occurring for this variable is found in the Pacific (25%), where the decrease is 1.31 (m/s)/century.
None of the trends in GHD are statistically significant (17% < P < 93%). The trends are decreasing in Asia, the Pacific and Canada. Europe and the Atlantic show an increase. The increase in the Atlantic is much higher than the decrease in other areas. The entire area shows an increase in GHD of 6.8 m/century. GH in the 30-40°N zone is increasing significantly in all areas, except the Pacific. The entire area shows a 33 m/century increase. The Atlantic area shows the strongest increase of 47 m/century. The Pacific has an increase of 24 m/century and is the area with the lowest increase, but also the highest P-value (6.7%). The 60-70°N zone differs from the 30-40°N zone. The GH here is also increasing, however not as fast as in its lower counterpart. The entire area shows a 27 m/century increase with a P-value of 0.5%. Canada has a similar P-value and stronger increase, while Europe and the Atlantic show smaller trends and have insignificant P-values. Asia and the Pacific have values similar to the entire area and are significant, however not strongly.
17
Correlations
Annual correlations are shown in tables 6 and 7 and are based on annual averages. Seasonal correlations are shown in tables 8 and 9 and based on seasonal averages. Tables 6 to 9 show the correlations of RWA and GHD with several variables. RWA is correlated to temperature, precipitation, GH and GHD. GHD is correlated to temperature and precipitation. In addition to the correlation coefficient, the P-values of the correlations are also shown in the tables.
Annual Correlations
Table 6 Correlation coefficients (left) and significance (right) of annual RWA (50-60°N) with
temperature (50-60°N), precipitation (50-60°N), GH (30-40°N & 60-70°N) and GHD (30-70°N).
Red numbers indicate negative correlations. The definition of the areas and abbreviations used
can be found in Tables 2 and 3. Green marks a significance of P < 1% and yellow marks a
significance of 1% < P < 5%.
Correlation and Significance Annual Rossby Wave Activity (50-60°N)
with other Variables
Variable Temperature (50-60N°) Precipitation (50-60°N) Geopotential Height (30-40°N) Geopotential Height (60-70°N) Geopotential Height Difference Ar e a World -0.20 17% -0.46 0.1% -0.63 0.0% 0.55 0.0% -0.85 0.0% Europe 0.00 100% -0.48 0.0% -0.53 0.0% 0.45 0.1% -0.64 0.0% Asia -0.15 28% -0.34 1.7% -0.42 0.2% 0.62 0.0% -0.73 0.0% Pacific 0.41 0.3% -0.15 29% -0.66 0.0% 0.65 0.0% -0.81 0.0% Canada 0.21 15% 0.14 34% -0.38 0.7% 0.48 0.0% -0.61 0.0% Atlantic 0.28 5.1% -0.42 0.2% -0.66 0.0% 0.73 0.0% -0.81 0.0%
Table 7 Correlation coefficients (left) and significance (right) of annual GHD (30-70°N) with
temperature (50-60°N) and precipitation (50-60°N). Red numbers indicate negative correlations.
The definition of the areas and abbreviations used can be found in Tables 2 and 3 (Green: P
<1%).
Correlation and Significance Annual
Geopotential Height Difference (30-70°N)
with other Variables
Variable Temperature (50-60N°) Precipitation (50-60°N) Ar e a World 0.17 25% 0.48 0.0% Europe 0.08 60% 0.58 0.0% Asia -0.02 86% 0.38 1.0% Pacific -0.55 0.0% 0.27 5.6% Canada -0.43 0.2% 0.14 32% Atlantic -0.51 0.0% 0.52 0.0%
18
Seasonal RWA
Table 8 Seasonal correlation coefficients (left) and significance (right) of RWA (50-60°N) with
temperature (50-60°N), precipitation (50-60°N), GH (30-40°N & 60-70°N) and GHD (30-70°N).
Red numbers indicate negative correlations. The definition of the areas and abbreviations used
can be found in Tables 2 and 3 (Green: P <1%. Yellow: 1% < P <5%).
World 0.27 5.5% -0.41 0.3% -0.45 0.1% 0.78 0.0% -0.80 0.0% Europe 0.03 84% -0.42 0.2% -0.54 0.0% 0.51 0.0% -0.61 0.0% Asia 0.16 28% 0.10 48% -0.38 0.6% 0.56 0.0% -0.59 0.0% Pacific 0.67 0.0% -0.25 8.5% -0.58 0.0% 0.84 0.0% -0.81 0.0% Canada 0.49 0.0% 0.06 68% -0.28 4.8% 0.71 0.0% -0.68 0.0% Atlantic 0.32 2.3% -0.26 7.1% -0.49 0.0% 0.71 0.0% -0.71 0.0% World -0.04 76% -0.22 13% -0.38 0.7% 0.29 3.8% -0.58 0.0% Europe 0.25 8.0% -0.32 2.4% -0.43 0.2% 0.59 0.0% -0.69 0.0% Asia -0.03 85% -0.00 99% -0.32 2.3% 0.61 0.0% -0.69 0.0% Pacific 0.09 52% -0.21 14% -0.42 0.2% 0.44 0.1% -0.60 0.0% Canada 0.05 72% -0.04 76% -0.25 8.5% 0.39 0.6% -0.48 0.0% Atlantic 0.27 5.3% -0.06 66% -0.32 2.4% 0.56 0.0% -0.65 0.0% World -0.17 25% -0.36 1.0% -0.48 0.0% 0.49 0.0% -0.70 0.0% Europe 0.04 78% -0.33 1.9% -0.16 26% 0.43 0.2% -0.48 0.0% Asia -0.04 76% -0.40 0.4% -0.42 0.2% 0.75 0.0% -0.80 0.0% Pacific 0.38 0.7% -0.07 63% -0.53 0.0% 0.69 0.0% -0.74 0.0% Canada -0.02 91% -0.24 9.3% -0.55 0.0% 0.57 0.0% -0.74 0.0% Atlantic 0.27 5.5% -0.14 32% -0.54 0.0% 0.64 0.0% -0.71 0.0% World -0.38 0.6% -0.64 0.0% -0.85 0.0% 0.79 0.0% -0.90 0.0% Europe -0.37 0.8% -0.71 0.0% -0.73 0.0% 0.50 0.0% -0.67 0.0% Asia -0.37 0.9% -0.58 0.0% -0.54 0.0% 0.70 0.0% -0.75 0.0% Pacific 0.58 0.0% -0.29 4.1% -0.70 0.0% 0.83 0.0% -0.85 0.0% Canada 0.22 13% 0.02 91% -0.58 0.0% 0.69 0.0% -0.72 0.0% Atlantic 0.41 0.3% -0.75 0.0% -0.85 0.0% 0.83 0.0% -0.88 0.0% A re a
Correlation and Significance Seasonal Rossby Wave Activity (50-60°N)
with other Variables
A re a
Winter
Variable Temperature (50-60N°) Precipitation (50-60°N) Geopotential Height (30-40°N) Geopotential Height (60-70°N) Geopotential Height DifferenceAutumn
Variable Temperature (50-60N°) Precipitation (50-60°N) Geopotential Height (30-40°N) Geopotential Height (60-70°N) Geopotential Height DifferenceSpring
A re a A re a Temperature (50-60N°) Variable Variable Precipitation (50-60°N) Geopotential Height (30-40°N) Geopotential Height (60-70°N) Geopotential Height DifferenceSummer
Temperature (50-60N°) Precipitation (50-60°N) Geopotential Height (30-40°N) Geopotential Height (60-70°N) Geopotential Height Difference19
Seasonal GHD
Table 9 Seasonal correlation coefficients and significance of GHD (30-70°N) with temperature
(50-60°N) and precipitation (50-60°N). Red numbers indicate negative correlations. The
definition of the areas and abbreviations used can be found in Tables 2 and 3 (Green: P <1%.
Yellow: 1% < P <5%).
Correlation and Significance Seasonal Geopotential Height Difference
with Temperature (50-60°N) and Precipitation (50-60°N)
Spring
Summer
Variable Temperature (50-60N°) Precipitation (50-60°N) Variable Temperature (50-60N°) Precipitation (50-60°N) Ar ea World -0.20 16% 0.55 0.0% Ar ea World -0.28 5.2% 0.14 35% Europe 0.21 14% 0.76 0.0% Europe -0.57 0.0% 0.40 0.4% Asia -0.36 1.0% 0.06 69% Asia -0.40 0.4% 0.07 65% Pacific -0.73 0.0% 0.46 0.1% Pacific -0.25 8.3% 0.34 1.5% Canada -0.63 0.0% 0.07 61% Canada -0.10 50% 0.28 4.9% Atlantic -0.51 0.0% 0.61 0.0% Atlantic -0.42 0.2% 0.10 51%Autumn
Winter
Variable Temperature (50-60N°) Precipitation (50-60°N) Variable Temperature (50-60N°) Precipitation (50-60°N) Ar e a World 0.01 94% 0.34 1.6% Ar e a World 0.32 2.4% 0.65 0.0% Europe -0.18 21% 0.47 0.1% Europe 0.35 1.3% 0.73 0.0% Asia -0.01 92% 0.46 0.1% Asia 0.02 88% 0.55 0.0% Pacific -0.48 0.0% 0.30 3.2% Pacific -0.59 0.0% 0.50 0.0% Canada -0.29 4.0% 0.30 3.6% Canada -0.41 0.3% 0.22 12% Atlantic -0.51 0.0% 0.34 1.5% Atlantic -0.63 0.0% 0.86 0.0%The annual correlations shown in table 6 indicate that annual mean RWA is not related to annual mean temperature. Precipitation shows a significant negative correlation with RWA in most areas but the significance in the Pacific and Canada is low. The significance for both GH zones and the GHD is very high. The upper GH zone shows positive correlations with RWA, while the lower zone and the GHD are negatively correlated to RWA.
The annual correlation of GHD with temperature and precipitation (table 7) are spatially fragmented. Between GHD and temperature, the low correlation and significance in Europe and Asia trump the negative correlations with high significance in the other areas. This leads to an insignificant correlation between GHD and temperature for the entire area. The significance in the correlation of GHD with precipitation is lacking in Canada and missing the mark by just 0.6% in the Pacific, but because all correlations between GHD and precipitation are positive this does not change the overall positive and significant correlation.
20
Seasonal correlations and their significance are shown in tables 8 and 9. Table 8 is examined first. This shows the seasonal correlations between RWA and temperature, precipitation, GH and GHD. The variables are highlighted individually.
For different seasons RWA correlates differently with temperature. Significances also depend on the season. In spring, all correlations are positive and three areas are significant. The entire area is almost significant (P=5.5%). In summer, there are no significant correlations between RWA and temperature and the values of the correlations are much lower than in spring. Autumn shows only the Pacific as significant, with a correlation between RWA and temperature four times stronger than in summer. The other correlations between RWA and temperature in autumn are no different from the summer ones. Winter shows significant correlations between RWA and temperature in almost all regions. World, Europe and Asia are negative while the Pacific, Canada and the Atlantic are positive.
A similar pattern is seen for the correlation between RWA and precipitation, where the significances are lowest in summer, somewhat present in spring and autumn but highly significant in winter. The
correlations in summer and autumn are all negative. In Spring, Asia and Canada have small positive correlations between RWA and precipitation and these correlations are not significant. Winter shows strong negative correlations with high significance in every zone except Canada. The winter correlation between RWA and precipitation in Canada is positive, but it is too small to be considered significant. The correlation between RWA and GH at 30-40°N shows strong negative correlations with high
significance in all seasons and areas. The only exceptions are Canada in summer (P=8.5%), and Europe in autumn (P=26%). Both insignificant correlations show correlation coefficients that are lower than all others. The correlations are strongest in Winter.
The correlation between the RWA and GH at 60-70°N has similar significance in correlations, except that Canada in summer and Europe in autumn are also significant. All correlations are positive, and again the correlations are strongest in winter, although on par to the ones in Spring.
The correlation between RWA and GHD has the same significance in all seasons and areas as the GH at 6070°N. Stronger and negative correlations can be seen. In winter, the correlation coefficient reaches -0.90 for the entire area.
Table 9 has the same structure as table 8. The GHD is correlated to temperature and precipitation. GHD is correlated differently to temperature per season. Spring shows an overall negative correlation with four out of five areas significant, with Europe having a P-value of 14%. In summer Europe shows the strongest negative correlation between GHD and temperature of all areas and this correlation is highly significant, while the correlation in Canada is lower, along with its significance. In autumn the overall significance is lower, whereas in summer the significance is higher again. It is only in winter that positive and significant correlations between GHD and temperature are seen, for the entire area and for Europe. GHD is positively correlated to precipitation in every season. In autumn the amount of areas with significant correlations between GHD and precipitation is highest. Summer has the least amount of areas with significant correlations. In general, the correlation coefficients of GHD and precipitation are highest in winter. For the values that are not significant, the correlations are positive but smaller than the others.
21
6. Discussion
The creation of the RWA parameter is discussed. The results, trends and correlations are interpreted and compared to other literature. The type of data used for this research is reviewed.
Recommendations for future research on the topic and suggestions for improvements to the Matlab (2012) script are found at the end of this discussion.
Rossby Wave Activity
Brandts (2015) defined Rossby Wave Activity (RWA) as the ratio of the absolute value of the meridional component of the wind speed over the zonal component:
𝑅𝑊𝐴𝐵𝑟𝑎𝑛𝑑𝑡𝑠= |𝑉|
𝑈 [-] (5)
The results of correlations calculated using this definition were insignificant and arbitrary. Dividing by the U-component means there is a possibility of dividing by zero, or values close to zero. This lead to noise amplification and infinitely large or small RWA values. When doing the analysis with this
definition, the U-component was not expected to show many negative values while in practice it does. More than 18% of the daily U-component values are negative.
Replacing the values of the U-component that were close to or below zero with positive values showed an increase in correlation values and significance. The large amount of negative values however made this an impractical option.
Testing different methods of calculating RWA led to the following definition, where the arctangent is used to obtain values representing wind direction:
𝑅𝑊𝐴𝑑𝑖𝑟= 0.5 −
𝐴𝑡𝑎𝑛 ( 𝑈
|𝑉|+0.001)
𝜋 [-] (6)
This definition used the direction of the Rossby wave as indication for its activity with values between 0 and 1. An airflow from west to east being valued as 0, an airflow from east to west valued as 1. This definition provides a scale where any deviation from a standard westerly is considered RWA. Two flaws are encountered. Discontinuities in RWA occur when calculating with wind speeds of near 0 m/s and the weight of the wind speed is lacking. A wind flowing from south to north with a certain speed should not be measured as equal to a flow in the same direction with double the speed. To rectify these flaws, the definition is multiplied with wind speed. The resulting and final definition is:
𝑅𝑊𝐴 = √𝑈2+ 𝑉2∗ (0.5 −𝐴𝑡𝑎𝑛 (
𝑈 |𝑉|+0.001)
𝜋 ) [m/s] (3)
Despite this definition yielding significant results, improvement is still possible. Tailleux and McWilliams (2000) for example investigate the influence of topography on Rossby waves, and Dickinson (1968) discusses the vertical movement of Rossby Waves. It might prove useful to consider that more variables are required to establish the best way of calculating RWA. Possible outcomes of equations 5, 6 and 3 are visualized in ‘Appendix IV: RWA Definitions’.
Trends
Temperature and precipitation both show a significant positive trend (Table 4 and 5; Figure 8). This concurs with the IPCC (2013b). The spatial variety in trends mentioned by Seneviratne (2016) is also
22
confirmed. Trends in temperature are stronger above continents than above oceans. The lack of significance for the precipitation trend in Europe, the Pacific and Canada could be related to seasonal differences in trends. For example, Xoplaki et al. (2000) show a decrease in precipitation during winter over Greece. This does not concur with the positive trend in precipitation over Europe shown in the results. However, a negative trend in winter might explain the P-value of 9.8%.
The distribution in which GH increases does not follow expectations. Despite the Arctic regions warming faster than the sub-tropics (Vaughan et al., 2003), according to the results of this research the GH increases faster in the latter. The expected decrease in GHD is not present (Table 4), the opposite is true: GHD between the sub-tropics and sub-Arctic regions appears to be increasing. Unfortunately, the resulting trend lacks the significance to be conclusive. Wang et al. (2017) show a positive trend in warming events in the Arctic troposphere. These warming events are fuelled by energy originating in the tropics. This indicates a relation between GH increase in the Arctic and GH increase in the tropics. The GH in the tropics being a strong influencing factor on the GH in the Arctic regions is a possible
explanation for the increase in GHD.
The effect of this unexpected increase in GHD on the meridional component of wind speed at 500 hPa varies. The V-component decreases in all areas, apart from the Atlantic. This could be related to the North Atlantic Oscillation (NAO), a geopotential height pattern or teleconnection (Wallace & Gutzler, 1981). According to Wallace and Gutzler (1981) the NAO is part of a global seesaw of geopotential height. The NAO could be responsible for negating the effect of decreasing RWA for the Atlantic. Related to this explanation is climate change forcing the NOA into an extreme phase (Hoerling et al., 2001), which would also mean an increase in meridional wind speeds above the Atlantic.
The U-component shows fluctuating trends, compared to the mostly negative trends for the V-component. This could be associated to the increase in GHD. Although all trends in GHD and the U-component are insignificant, the two variables might be correlated. Both variables show a positive trend for Europe and the Atlantic. This partially concurs with Hurrell (1995), who claims that positive trends in NAO result in stronger westerlies in Canada, the Atlantic and Europe. The Pacific and Canada show a negative trend. The lack of significance prevents the conclusion that a decreasing GHD leads to a decrease in zonal wind speed, which allows for a stronger RWA.
Overall RWA shows a slight decrease, but these trends are not significant. The lack of any trend in Europe is an exception. The size of the trends is strongest over the Atlantic and weakest over the Pacific. The values being all negative could be an indication that in practice RWA is decreasing, but this is not scientifically established yet. Looking at a longer period might increase the significance. Due to the relatively recent nature of climate change however, using a longer period to look at RWA would likely result in even lower trends. Lower trends would in turn provide lower P-values.
According to Overland et al. (2011) the increase in GH over the Arctic is caused by the loss of sea ice. The positive trend calculated in this research confirms a significant increase in GH. A lower trend in the tropics and thus a decreasing difference in GH however is not mentioned in their research, instead they show lower GHs in lower latitudes. A possible explanation could be related to trends being different at each pressure level, since Overland et al. (2011) focus more on the 850 hPa plain. As Marshall (2002) indicates, trends differ when looking at different pressure levels.
23
A second explanation for the GH increasing faster in the Tropics is that the oceans are significantly reducing the average increase. The GH in the Northern Pacific between 1977 and 1986 was extremely low (Nitta & Yamada, 1989). An extreme low GH lasting nearly 10 years would have a strong impact on the data, considering the data used in this research takes only 50 years into account. This leads to underestimations of trends in GHD. Although the Pacific does show a strong positive trend, in the 60-70°N area the Atlantic does not. The trends calculated in this research do not concur with Nitta & Yamada (1989).
A third explanation for the increase in GHD is that polar amplification is limited to latitudes higher than 60-70°N. Ice-albedo plays a key role in polar amplification (Holland & Bitz, 2003). A stronger, more significant negative trend in GHD might be found when calculating the difference using higher latitudes.
Correlations
The correlations of RWA with temperature and precipitation seem to be strongly related to the seasons (Table 8). Especially when looking at temperature, the annual correlation with RWA (Table 6) shows nearly no significance, while the seasonal correlations show otherwise. The correlation between RWA and temperature is highly significant in winter, while not significant at all in summer. This seasonal difference in significance and strength of correlation repeats itself for the other variables and is possibly related to the amplification of RWA in winter caused by troughs formed behind mountain ranges such as the Rocky Mountains and the Tibetan Plateau (Manabe & Terpstra, 1974). It is best seen in temperature and precipitation. It is also slightly present in the correlation of RWA with GH and GHD, however the fact that almost all those correlations are highly significant makes it more difficult to see.
Plumb (2010) also confirms strong RWA in winter. Plumb (2010) discusses interactivity between Rossby waves and the stratosphere due to a shared connection to the mean flow. Despite Plumb (2010)
focusing his research more on the stratosphere, the increased correlations between RWA and climate in winter could apply to the 500 hPa plain as well.
The correlation between RWA and GH and GHD (Tables 6 and 8) is higher and more significant than originally expected. This correlation was not present when using the RWA definition used by Brandts (2015). Both annual and seasonal correlations are almost all highly significant. Summer shows slightly higher P-values than other seasons for both GH zones, which are expectedly caused by the negation of troughs behind mountain ranges (Manabe & Terpstra, 1974). Correlations between RWA and GHD show a 0,0% P-value for all areas and all seasons. The negative correlation between RWA and GH with the 30-40°N zone indicates that RWA decreases when GH at 30-30-40°N increases. Considering the created RWA definition this means a decrease in meridional wind speeds and an increase in westerlies. Opposite is the positive correlation of RWA with the GH at 60-70°N.
RWA shows a negative correlation with GHD. This means that RWA decreases when the GHD increases. Even though an increase in GHD is not significantly proven in this research, a significant and stronger increase in the 30-40°N zone than in the 60-70°N zone indicate that the GHD might very well be increasing. That increase concurs with the overall negative trend in RWA, although not significant. In addition to correlating temperature and precipitation to RWA, GHD is also correlated to temperature and precipitation (Table 7). The world shows an insignificant and small positive correlation between GHD and temperature, whereas there is no correlation at all between GHD and temperature in Europe
24
and Asia. The Pacific, Canada and the Atlantic show highly significant (P<1%) negative correlations between GHD and temperature.
The areas having negative annual correlations between GHD and temperature have negative
correlations (mostly significant) in all seasons (Tables 7 and 9). In winter, where almost all correlations in this research show higher significances and stronger correlation coefficients, the correlation between GHD and temperature is positive for the world and Europe. This is unexpected, since except for Europe in spring, all correlations are otherwise either negative or near zero. Considering the values resulting from these other correlations however, these positive correlations between GHD and temperature not likely random. The winter correlation of RWA and temperature shows a mirrored image, positive correlations where table 9 shows negative ones and the other way around. This confirms the negative correlation between RWA and GHD.
The correlation of GHD with precipitation is positive (Tables 7 and 9). The strength of these correlations varies spatially between somewhat present and strongly present. The correlations between GHD and precipitation are strongest in winter. Autumn has more significant correlations than winter, due to Canada having a P-value of 12% in winter. The positive correlations of GHD with precipitation confirm the same as the negative correlations of GHD with temperature: RWA is negatively correlated to GHD.
Reanalysis Data
The type of data used for this research is reanalysis data. Reanalysis data consists of a mixture of modelled and measured data and therefore provides the major advantage of having a wide range of data availability. It allows for trend and correlation calculations as done in this research, where the focused area is a strip of 30 degrees in latitude and 360 degrees in longitude. There are several other advantages to using reanalysis data, mainly related to data coverage and accessibility (UCAR, 2016). There are also some downsides in using reanalysis data. The main downside experienced in this research is the size of the files. When obtaining the datasets from the ECMWF (2017), the downloads were interrupted or corrupted on several occasions due to their size. The size of the original downloaded files also proved challenging when imported in Matlab (2012). Despite having plenty of computing power, the software’s limit was reached when trying to process all downloaded files at the same time. Other downsides to using reanalysis data concern the data reliability. Small scale events or anomalies might be lost due to the interpolation used in creating reanalysis data. Software programs such as Panoply (2017) can help in finding possible discrepancies by creating simple plots per time step that allow for a quick oversight in downloaded reanalysis data with NetCDF extension. Areas with
unexpected or missing values are quickly noticed in these plots and can easily located. Fortunately, this research focusses on large areas, which have lower variations from reality than smaller sites (Rose & Apt, 2015). In general reanalysis data is suitable for this research, as long biases caused by using different models and types of measurements are sufficiently corrected (Bengtsson et al., 2004).
Recommendations
Further research on the topic to improve the understanding of how the relationship between Rossby Wave Activity and climatological trends works would benefit from using data of a longer time span. The 50 years focused on in this research proved to be sufficient in calculating significant trends and
correlations for most variables, but the significance of trends in Rossby Wave Activity is still too low to be conclusive. Performing the same calculations on a timespan of for example 60 years would increase
25
significances. A possibility is to use data from 1956 to 2016. With this timespan the data would benefit from 6 more years of recent climate change. Significances will increase when combining a longer time span with newer reanalysis data of higher quality (Dee et al., 2014). In addition to the time scale, the Southern Hemisphere is also a point of interest. The oversight provided by Ventusky (2017) quickly shows a much higher activity in wind speeds at the 500 hPa in the Southern Hemisphere, compared to the Northern Hemisphere.
Besides the timescale and area, the resolution of the used data is also a point of interest. Using the 10x10 degree fields proved to be effective in finding trends and correlations. The downside of creating areas that represent continents using 10x10 fields is that the boundaries are rough estimates and not exact boundaries. It would benefit the precision of the results to use lower resolution data to more accurately define boundaries of areas, instead of combining the 10x10 fields to do so.
The Matlab (2012) script created could be improved for performing similar analyses. The trend
calculations section for example is already largely automated. The large initial file size is an obstruction that prevents the full script from being able to run in one go. The organizing and saving of results is a second challenge. Improving the script to be run entirely in one go would be time consuming but worth it when looking at similar data sets of large size or quantity.
A final addition that might prove useful is to compare the other variables, besides GH, on several latitudes as well. This would greatly increase the total amount of correlations to be calculated. Considering the added value of correlating the GH at 30-40°N and 60-70°N however, the results could provide useful insights. An example of a research that focusses on temperature rise across different latitudes is Deutsch et al (2008). They show temperature increases faster along with latitude, while consequences are likely to be most severe in lower latitudes.
7. Conclusion
The aim of this research is to establish the influence of climate change on Rossby Wave Activity (RWA) in the Northern Hemisphere and consequently the effect thereof on climatological trends. In order
establish the correlation between RWA and other climate variables a new parameter was defined that can be used to describe RWA quantitatively. Using wind speed and direction to define RWA has proven to be valuable in looking for correlations with other variables.
Trends are found for temperature, precipitation and GH. Unexpected is the 30-40°N zone showing a more rapid increase in GH than the 60-70°N zone. For all these variables, the significance in trends is highly significant for the entire area. Some individual areas lack significance. Where there is a lack of significance, the calculated trend is smaller than the other zones.
Trends in the absolute V-component are mostly negative. The NAO is a probable cause for the Atlantic being an exception. The U-component has positive trends in some areas and negative trends in other areas, but none of these trends are statistically significant. The GHD shows results comparable to the U-component. Some areas have positive trends, others negative, but the significance is lacking. This is different from the trends in RWA. Despite the lack of significance, it is likely that RWA is decreasing since all areas show small negative trends, except for the missing trend in Europe.
Strong correlations are found between RWA and GHD, both on an annual and on a seasonal basis. The correlations of RWA with temperature and precipitation are different for every season, strongest in
26
winter and weakest in summer. An increasing GHD results in a decreasing RWA. An increase in GHD occurs due to a faster rise of GH in the sub-tropics than the sub-Arctic regions.
The GHD shows fragmented correlations with temperature. The correlation is mostly negative but on occasion positive and increases in significance during winter. GHD is positively correlated to
precipitation in all seasons and areas. The strength of the correlation increases in winter, the
significance however is highest in autumn. This exception is caused by a low correlation between GHD and precipitation during winter for Canada.
To answer to the main research question: ‘What is the relationship between Rossby waves at the 500 hPa plain and climatological trends at temperate latitudes on the Northern Hemisphere?’, the results of this research are not conclusive. Although the correlations between RWA and GH variables are highly significant, the trends in RWA and GHD are not. The relationship between RWA and climatological trends is present but because there is no significant trend in RWA it cannot be scientifically confirmed that this correlation is influencing climatological trends.
Possible recommendations for further research are: use a longer timespan to perform calculations, broaden the area of interest, increase the data resolution, optimize the Matlab script and include more variables.
References
Aguado, E. & Burt, J.E. (1999). Understanding Weather & Climate. Upper Saddle River, New Jersey, Pearson Prentice Hall.
Anita, W., Dominic, M. & Neil, A. (2010). Climate Change and Agriculture Impacts, Adaptation and Mitigation: Impacts, Adaptation and Mitigation. OECD publishing, ISBN 978-92-64-08687-6. Barlow, M., Cullen, H., & Lyon, B. (2002). Drought in central and southwest Asia: La Nina, the warm pool,
and Indian Ocean precipitation. Journal of climate, 15(7), 697-700.
Becker, E., Müllemann, A., Lübken, F. J., Körnich, H., Hoffmann, P., & Rapp, M. (2004). High Rossby‐wave activity in austral winter 2002: Modulation of the general circulation of the MLT during the MaCWAVE/MIDAS northern summer program. Geophysical research letters, 31(24).
Bengtsson, L., Hagemann, S. & Hodges, K. I. (2004). Can climate trends be calculated from reanalysis data? Journal of Geophysical Research: Atmospheres, 109(D11).
Biastoch, A., Treude, T., Rüpke, L. H., Riebesell, U., Roth, C., Burwicz, E. B., ... & Wallmann, K. (2011). Rising Arctic Ocean temperatures cause gas hydrate destabilization and ocean acidification. Geophysical Research Letters, 38(8), L08602.
Bony, S., Colman, R., Kattsov, V. M., Allan, R. P., Bretherton, C. S., Dufresne, J. L., ... & Randall, D. A. (2006). How well do we understand and evaluate climate change feedback processes? Journal of Climate, 19(15), 3445-3482.
Brandts, B. (2015). The Effects of Climate Change on the Upper Air Circulation over Temperate Latitudes. University of Amsterdam, Earth Science Master Thesis.
27
Cox, P. M., Betts, R. A., Jones, C. D., Spall, S. A. & Totterdell, I. J. (2000). Acceleration of global warming due to carbon-cycle feedbacks in a coupled climate model. Nature, 408(6809), 184-187. Dee, D. P., Balmaseda, M., Balsamo, G., Engelen, R., Simmons, A. J. & Thépaut, J. N. (2014). Toward a
consistent reanalysis of the climate system. Bulletin of the American Meteorological Society, 95(8), 1235-1248.
Deutsch, C. A., Tewksbury, J. J., Huey, R. B., Sheldon, K. S., Ghalambor, C. K., Haak, D. C. & Martin, P. R. (2008). Impacts of climate warming on terrestrial ectotherms across latitude. Proceedings of the National Academy of Sciences, 105(18), 6668-6672.
Dickinson, R. E. (1968). Planetary Rossby waves propagating vertically through weak westerly wind wave guides. Journal of the Atmospheric Sciences, 25(6), 984-1002.
Epstein, P. R. (2001). Climate change and emerging infectious diseases. Microbes and infection, 3(9), 747-754.
ECMWF (European Centre for Medium-Range Weather Forecasts) (2017). [Online] http://www.ecmwf.int , last reviewed on March 2017.
Francis, J. A. & Vavrus, S. J. (2012). Evidence linking Arctic amplification to extreme weather in mid‐ latitudes. Geophysical Research Letters, 39(6), L06801.
Hoerling, M. P., Hurrell, J. W., & Xu, T. (2001). Tropical origins for recent North Atlantic climate change. Science, 292(5514), 90-92.
Holland, M. M., & Bitz, C. M. (2003). Polar amplification of climate change in coupled models. Climate Dynamics, 21(3-4), 221-232.
Hurrell, J. W. (1995). Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science-AAAS-Weekly Paper Edition, 269(5224), 676-678.
IPCC, 2013a: Annex III: Glossary [Planton, S. (ed.)]. In: Climate Change 2013: The Physical Science Basis.
Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A.
Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
IPCC, 2013b: Summary for Policymakers. In: Climate Change 2013: The Physical Science Basis.
Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A.
Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.
Kelley, C. P., Mohtadi, S., Cane, M. A., Seager, R., & Kushnir, Y. (2015). Climate change in the Fertile Crescent and implications of the recent Syrian drought. Proceedings of the National Academy of Sciences, 112(11), 3241-3246.
28
Kreibich, H., Thieken, A. H., Petrow, T., Müller, M., & Merz, B. (2005). Flood loss reduction of private households due to building precautionary measures--lessons learned from the Elbe flood in August 2002. Natural Hazards and Earth System Science, 5(1), 117-126.
Lee, T. M., Markowitz, E. M., Howe, P. D., Ko, C. Y. & Leiserowitz, A. A. (2015). Predictors of public climate change awareness and risk perception around the world. Nature Climate Change, 5(11), 1014-1020.
Lillo, S. P., & Parsons, D. B. (2017). Investigating the dynamics of error growth in ECMWF medium‐range forecast busts. Quarterly Journal of the Royal Meteorological Society, 143(704), 1211-1226. Lutgens Frederick, K., & Tarbuck, E. J. (2001). The Atmosphere. ISBN 978-0130879578.
Manabe, S., & Terpstra, T. B. (1974). The effects of mountains on the general circulation of the
atmosphere as identified by numerical experiments. Journal of the atmospheric Sciences, 31(1), 3-42.
Marshall, G. J. (2002). Trends in Antarctic geopotential height and temperature: A comparison between radiosonde and NCEP–NCAR reanalysis data. Journal of Climate, 15(6), 659-674.
Martínez-del-Río, J. & Céspedes-Lorente, J. (2014). Competitiveness and legitimation: The logic of companies going green in geographical clusters. Journal of business ethics, 120(1), 131-146. MATLAB [Computer Software]. (2012). Matlab v2012b, The MathWorks, Inc., Natick, Massachusetts,
United States.
McMichael, A. J., Woodruff, R. E. & Hales, S. (2006). Climate change and human health: present and future risks. The Lancet, 367(9513), 859-869.
Nitta, T. & Yamada, S. (1989). Recent warming of tropical sea surface temperature and its relationship to the Northern Hemisphere circulation. Journal of the Meteorological Society of Japan. Ser. II, 67(3), 375-383.
NWS (National Weather Service) (2017). Absolute Vorticity, [Online]
http://www.srh.noaa.gov/jetstream/constant/vort_max.html , last reviewed on the 3rd of
October 2017.
Oetken, M., Stachel, B., Pfenninger, M., & Oehlmann, J. (2005). Impact of a flood disaster on sediment toxicity in a major river system–the Elbe flood 2002 as a case study. Environmental
Pollution, 134(1), 87-95.
Olesen, J. E., Trnka, M., Kersebaum, K. C., Skjelvåg, A. O., Seguin, B., Peltonen-Sainio, P., ... & Micale, F. (2011). Impacts and adaptation of European crop production systems to climate
change. European Journal of Agronomy, 34(2), 96-112.
Overland, J. E., Wood, K. R. & Wang, M. (2011). Warm Arctic—cold continents: climate impacts of the newly open Arctic Sea. Polar Research, 30(1), 15787.
Panoply [Computer Software] (2017). Panoply 4.7.0, Panoply NetCDF, HDF and GRIB Data Viewer. Retrieved from https://www.giss.nasa.gov/tools/panoply/ , last reviewed on the 21st of March
