Topic 1.1 Weather Data Annotated PDF

Dr Jessie Copper
School of Photovoltaic and Renewable Energy Engineering
SOLA9103 – RE System Modelling and Analysis
Topic 1.1: Weather Data Sources
The first topic in SOLA9103 is on weather data sources.

Weather data underpins all renewable energy system modelling. From resource
assessments, pre-feasibility studies, yield analysis, all the way to financial analyses and
performance guarantees.
No matter how accurate your underlying energy model is, if you utilise poor weather
data, you’ll achieve poor results from your model.

Topic 1.1: Weather Data
Dr. Jessie Copper 0
Learning Outcomes
By the end of this lecture, you should be able to:
• Identify available weather data sources;
• Describe the difference between various types of weather files; and
• Begin to understand why it is important to assess weather data for RE system
modelling.
School of Photovoltaics and Renewable Energy Engineering
By the end of this lecture you should be able to identify a variety of weather data
sources; describe the difference between various types of weather files; and begin to
understand why it is important to assess weather data for RE system modelling. In future
lectures we will explore the differences in these data sources in more detail.
This slide presents the logo’s from a handful of companies which provide weather data.

Dr. Jessie Copper 1
Lets Start with a Simple Question...
What is the long-term average value of global

horizontal irradiation (GHI) for our current
location of UNSW, Sydney, Australia?
To begin, let’s start with what looks like a relatively simple question to answer.
What I want to know is what is the long-term average value of global horizontal
irradiation (GHI) for our current location of UNSW, Sydney, Australia?
Now, you don’t need to know this value off the top of your head, but what I do want you
to think about is, where would you go to find the answer to this question? What source
of data would you refer to obtain the answer? Would you refer to multiple data sources?
Keep these questions in mind as we progress through this topic.

Dr. Jessie Copper 2
mean = 4.6 kWh/m2.day
How Would You Respond?

range = 0.4 kWh/m2.day or 8.9%
std = 0.12 kWh/m2.day or 2.7%
(excluding MERRA-2 & IWEC2)
Wait, I looked up
That’s easy, just IRENA’s Global
I looked up NASA But, I looked up the World Bank What about TMY
let me look up the Atlas for And MERRA-2
POWER and Meteonorm and Group’s Global Weather files?
BoM website, Renewable says the value is..
it’s... the value is... Solar Atlas and
it’s ... the value is... Energy says it’s..
4.6 4.45 4.59 4.54 4.51 5.2 ERMY – 4.62

kWh/m2.day kWh/m2.day kWh/m2.day kWh/m2.day kWh/m2.day kWh/m2.day IWEC – 4.53
IWEC2 – 4.18
RMY – 4.86
kWh/m2.day
So, how did you respond to the question?

This slide reveals that there exists a number of ways you could have responded. This
slide lists 10 different sources of data that you could have referred too in order to obtain
an answer. NB: This list is by no means exhaustive, other data sources also exist.
In this module we will explore some of these weather data sources, but to put this
example into the broader context of SOLA9103 do you happen to notice anything in
particular about the 10 answers provided on this slide?
For instance, did you happen to notice that the values from the 10 data sources are not
identical. You should observed that the range across the 10 data sources is
approximately 1 kWh/m2.day, or in relative terms the range is 22% of the mean value of
4.6 kWh/m2.day. In this simple example, we can also identify that MERRA2 and IWEC2
are outliers, when viewed in comparison to the other data sources. Excluding these two
data sources reduces the range down to a relative value of 9%. However, even after
excluding the two outlying data sources, you might still be asking what is the long-term
average value of GHI at UNSW?
At this point you might have developed some additional questions. For instance, which
of the 10 data sources is correct? Should we refer to more than one data source? Should
we provide the average across all the data sources? Do we need to report a level of
uncertainty around the answer we provide?.
We will aim to answer these and many more questions throughout this course, but for
now, we just want to begin by familiarising ourselves with some of these data sources.

Dr. Jessie Copper 3
Types of Data for RE Modelling
• Geospatial vs. Temporal data (or Gridded vs. Point data)
• Geospatial data – commonly referred to as raster or gridded data
• Represents the relative position of a variable on the Earth’s surface
• Temporal or Discrete Point location data
• Represents a variable at a specific location across a specified time period
vs.
Source: S. Cox et al., Renewable Energy Data, Analysis and Decisions: A Guide for Practitioners, March 2018, NREL/TP-6A-20-68913,
https://www.nrel.gov/docs/fy18osti/68913.pdf
When we talk about weather data sets, and more broadly renewable energy data sets, it
is important to understand what we mean by the terms geospatial, temporal, measured,
modelled and validated data. Hence, let’s define what we mean by each of those terms.
The following has been extracted NREL’s technical report: S. Cox et al., Renewable
Energy Data, Analysis, and Decisions: A Guide for Practitioners, March 2018, NREL/TP-
6A20-68913,
Geospatial data describe the relative position of something on the Earth’s surface.
Examples include the location of roads or cities or the amount of renewable energy
resource at a specific location.
Temporal data have specified dates or times. Temporal data can be discrete in time;
temporal data are assumed to contain repeated observations—for example, wind speed
measurements at 5-minute intervals.
Continuous gridded data, commonly referred to as raster data, are distributed evenly
across the area of interest. For example, the area might be divided into 2 km2 by 2 km2,
and a data point at the centre of each square represents an average or total value for
the entire square’s area. Continuous data are generally modelled or interpolated.
Discrete point location data, commonly referred to as vector or point data, are typically
not uniform and represent the location of ground measurement stations. Sometimes
these stations have additional meteorological information that can be very useful;
however, the stations are likely to be widely dispersed.

Dr. Jessie Copper 4
Types of Data for RE Modelling
• Modelled, Measured and Validated Data
• Modelled data – can be geospatial or temporal data
• Measured data – point specific data, useful for site specific assessment
• Validated data – modelled data that has been compared to measured data
Source: S. Cox et al., Renewable Energy Data, Analysis and Decisions: A Guide for Practitioners, March 2018, NREL/TP-6A-20-68913,
The following has been extracted from NREL’s technical report: S. Cox et al., Renewable
Energy Data, Analysis, and Decisions: A Guide for Practitioners, March 2018, NREL/TP-
6A20-68913,
Modelled data are renewable energy resource data developed with modelling
methodologies. For example, atmospheric scientist use a variety of methods to estimate
solar irradiance, including empirical, semiempirical, and more recently, physics-based
methods. For wind resource data, atmospheric scientists user numeric weather models
(e.g.. Weather Research and Forecasting Model) and downscaled reanalysis datasets.
Measured data are point-specific renewable energy data gathered by trained personnel
or meteorological measuring equipment. If many points are available across a region,
then these data can provide a reasonable understanding of the solar resource for a
region. These data can also be used for site-specific project assessments.
Validation of data are typically performed at the time renewable energy resource data
are created. Validation refers to comparing modelled data to “ground truth” data to
quantify how closely the model matches reality. Validation can take many forms – for
example, from a meteorological standpoint, comparing diurnal and annual cycles of wind
speed, or from a power standpoint, comparing power output ramps and frequency
distributions. Variation can play an important role in analysis application, sometimes
warranting the use or purchase of another dataset if the validation metrics prove
systematic biases.

Dr. Jessie Copper 5
Sources of Weather Data

Dr. Jessie Copper 6
Australian Bureau of Meteorology
• www.bom.gov.au
• Network of Automatic Weather Stations (AWS)
– Over 550 stations monitoring temperature,
humidity, wind speed and direction, pressure,
rainfall and cloud cover
– Approx. 200 stations with more than 20 years of
temperature/wind speed measurements
A Temporal and Point location dataset
Let’s now take a look at some specific sources of data.

The primary source of weather data in Australia is from the Australian Bureau of
Meteorology. Otherwise referred to as the BoM. Other countries have similar
meteorological offices for example the “Met Office” in the United Kingdom and the
Japan Meteorological Agency in Japan. This lecture will primarily focus on the weather
data source available across Australia, however we will also look at some global data
sources that are available.
The Australian BoM operates a network of over 550 AWS across Australia, as shown in
the figure. This network of stations is designed to serve the dual purpose of providing
real-time data for the Bureau’s forecasting, warning and information services, as well as
providing high quality data for the Bureau’s climate database.
This network of stations measures the standard weather variables like temperature,
humidity, rainfall and wind speed, and the data is available on a variety of time intervals
including minute, half-hourly, 3 hourly, Daily and Monthly. It should be noted that the
weather variables measured at each station varies, as does the historical period of
record and the completeness of data.
For example, in Sydney the AWS located at Observatory Hill does not measure wind
speed or direction. These variables are however measured at the nearby station located
at Fort Denison. Together these two stations are used to report the weather for the city
of Sydney.
The latest observations from a selection of the BoM AWS are available for the past 72
hours. For example, the latest weather observations for the AWS across NSW can be
found here http://www.bom.gov.au/nsw/observations/nswall.shtml
Finally, it is important to note that this network of AWS does not measure solar
irradiance.

Dr. Jessie Copper 7
• One minute solar irradiance data
– http://www.bom.gov.au/climate/data/oneminsolar/about-IDCJAC0022.shtml
– 21 locations available
– 11 stations currently in operation
» 3 stations in WA restarted in 2017
– Various years of operational data available
In addition to the Bureau’s network of AWS, the BoM also owns and operates a network
of solar irradiance stations with data recorded at a 1 minute resolution. Historical data is
available from 21 locations across the country, as shown in the figure. However, only 11
of these 21 stations are currently in operation, and three of the stations, all located
across Western Australia, were only reinitiated at the beginning of 2017. As per the
network of AWS, the data availability and completeness of the solar irradiance data
varies from station to station.

Dr. Jessie Copper 8
• Climate Data Online
– http://reg.bom.gov.au/climate/data/
– Daily and Monthly Statistics
» Rainfall
» Temperature
» Solar Exposure (satellite derived)
– Search by:
» Station number
» Location name or latitude and longitude
In addition to the raw data, the BoM also provides daily and monthly statistics for each
AWS. This data is made freely available via the BoM website. www.bom.gov.au
The data is searchable via station number, location name or via latitude and longitude
coordinates. Statistics are available for the three parameters of Rainfall, Temperature
and Solar Exposure; where solar exposure is equivalent to global horizontal insolation in
kWh/m2.day and is derived from satellite imagery i.e. a modelled data set.

Dr. Jessie Copper 9
• Gridded Spatial Data
– Solar Energy
» Daily solar exposure
» Monthly and annual solar exposure
» Monthly average hourly global and direct normal solar
exposure
» Hourly direct normal and global solar irradiance
– Maps – Recent conditions http://www.bom.gov.au/climate/maps/

» Rainfall, Temperature, Humidity & Solar Exposure
A Geospatial or Gridded dataset
Finally, the Bureau also offers a range of data in gridded format for use in spatial
modelling. The most important for PV system modelling is the hourly direct normal and
global horizontal irradiance dataset. This data set consists of 26+ years of hourly data
from 1990 through to near real time, and is available at a spatial resolution of 0.05
degrees, which is approximately equivalent to a grid of 5 by 5 kms.
It should be noted that this dataset is derived from satellite imagery and is a modelled
dataset.
Other gridded datasets available include daily and monthly averages of mean, max and
min air temperatures, rainfall and humidity. The example presented in this slide is a loop
of daily maximum temperatures in January 2017, on the introductory slide we saw the
loop of daily minimum temperatures for January 2017.

Dr. Jessie Copper 10
AREMI – National Maps
– https://nationalmap.gov.au/renewables/
– A spatial data platform for the Australian Energy Industry
– Includes:
» Annual & monthly climatology's BoM satellite GHI & DNI
» Hourly GHI & DNI (1990 – 2015 both years inclusive)
» Solar stations hourly historical observations – Created from BoM 1 minute solar and AWS
data
» TMY weather files – Created from BoM 1 minute solar and AWS data
» Wind speed map @ 100m and 150m
» Electricity Infrastructure and more... A Geospatial or Gridded dataset
The Australian Renewable Energy Mapping Infrastructure, otherwise known as AREMI,

provides access to renewable energy and other general information which has been
provided by various third party data custodians.
Some of the data layers of interest for renewable energy modelling, include:
• The BoM annual and monthly satellite derived Global Horizontal Irradiance (GHI) and
Direct Normal Irradiance (DNI) climatologies,
• The BoM hourly satellite derived GHI & DNI data from 1990 to 2015, downloadable
as a CSV file for a selected location
• The hourly historical observations and TMY weather files generated by AREMI from
the BoM 1 minute solar irradiance and AWS data
• Wind speed maps at heights of 100 and 150m from DNV GL.
Other general information available on the map include elevation and information about
the electrical infrastructure.

This slide presents two data streams that are viewable on AREMI. These are the BoM
annual and monthly satellite derived GHI & DNI climatologies (left), the hourly historical
observations & TMY weather files generated by AREMI from the BoM 1 minute solar
irradiance and AWS data (right).

Also available on AREMI to visualise and download is the BoM hourly GHI & DNI (1990 –
2015 both years inclusive).
This video plots the BoM hourly DNI for a three day period from 01/12/2014 to
03/12/2014, both days inclusive. The blue regions indicate areas of high cloud cover and
lower horizontal irradiance. The red regions indicate areas with low to no cloud cover
and high levels of irradiance. A visualisation of the full year of data for 2014 is available
on the AREMI platform. https://nationalmap.gov.au/renewables/

AREMI also has available wind speed maps at heights of 100m (left) and 150m (right)
from DNV GL.

Other general information available on AREMI include, maps of elevation (left and right)
and information about the electrical infrastructure (right).

NASA
– https://power.larc.nasa.gov/
– Daily, monthly or annual time series from July 1983 through near real time
– Spatial resolution 1 by 1 or 0.5  by 0.625 
– Parameters include:
» Solar geometry
» Insolation
» Temperature
» Wind speed and direction
– Direct access to the data via PV modelling software PVsyst
» Monthly averages of temperature and insolation
In terms of datasets Worldwide, NASA offers a series of daily, monthly and annual
average gridded datasets and statistics.
In comparison to the BoM, NASA’s gridded datasets and statistics are published at a
significantly lower resolution i.e. 1 or 0.5 vs. 0.05.
This has an impact on our ability to assess spatial variability or assess differences in the
meteorological variables between two sites. It also has an impact on the uncertainty of
the estimate for a location.
For example, if we are looking at two locations 50km apart, we are likely to see no
difference in the monthly statistics using the NASA dataset, but we would see a
differences if we used the BoM gridded dataset, simply because of the datasets spatial
resolutions.

So why then do I highlight the NASA dataset?

Well, this dataset can be directly accessed through the PV modelling package, PVsyst.
Hence, it is really simple to automatically import the long term historical averages of
insolation, air temperature and wind speed into PVsyst for a new site, but before doing
so, we need to consider the spatial resolution of this dataset, and consider whether this
is the most accurate method for your modelling task.

NASA
• MERRA & MERRA-2 reanalysis gridded dataset
– https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/
– 39+ years of data (1980 – current)
– Spatial resolution 0.5 x 0.625
– Various time intervals available i.e. monthly, 3-hourly, hourly
– 72 vertical levels of data available
– netCDF data format
• Other reanalysis datasets available worldwide include:

– NCEP-NCAR, NCEP-DOE, CFSR, ERA5 & JRA-55 A Geospatial or Gridded dataset
Another global gridded dataset offered by NASA is the MERRA and MERRA2 reanalysis
datasets.
A reanalysis dataset gives a numerical description of the recent climate, produced by
combining spatial/global models with point based observations. Such datasets contain
estimates of atmospheric parameters such as air temperature, pressure and wind
speeds and directions at different altitudes and surface parameters such as rainfall, soil
moisture content, and sea-surface temperatures. These estimates are produced for all
locations on the earth, at various time resolutions and they span a long time period.
MERRA is just one of the available reanalysis datasets, other examples include NCEP-
NCAR, CFSR, ERA-Interim and JRA-55. The advantage of the MERRA/MERRA2 reanalysis
dataset is the availability of the data at an hourly resolution. Other reanalysis datasets
currently have minimum temporal resolutions on the order of 6 and 3 hourly, making
them less suitable for hourly simulations.
The disadvantage of the MERRA dataset is its spatial resolution of 0.5 x 0.625 i.e. a
grid of approx. 50 x 63 km. Conversely, the advantages of the dataset is its length of the
historical data, and the availability of hourly grids of temperature and wind speed. Such
hourly grids of these parameters are not currently available from the BoM, only daily
and monthly grids of temperature are available.

ERA5
• ERA5 from ECMWF (European Centre for Medium Range Weather Forecasts)
– https://www.ecmwf.int/en/research/climate-reanalysis
– https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-single-
levels?tab=overview
– 40+ years of data (1979 – current)
– Spatial resolution 0.25 x 0.25
– Hourly resolution
– GRIB or netCDF data format
Another hourly reanalysis dataset is the ERA5 reanalysis dataset from ECMWF (European
Centre for Medium Range Weather Forecasts). This reanalysis dataset is also available
globally on an hourly time resolution, and includes all the standard variables like 10m
wind speed & direction; clear-sky or downward surface radiation (i.e. GHI); 2m dew
point and air temperature; and total precipitation (i.e. rainfall).
What is climate reanalysis? by ECMWF ‘A climate reanalysis gives a numerical

description of the recent climate, produced by combining models with observations. It
contains estimates of atmospheric parameters such as temperature, pressure and wind
at different altitudes and surface parameters such as rainfall, soil moisture content, and
sea-surface temperatures. The estimates are produced for all locations on earth, and
they span a long time period that can extend back by decades or more. Climate
reanalyses generate large datasets that can take up several petabytes of space.’

NCEI - Integrated Surface Database (ISD)
• NCEI – National Centres for Environmental Information
• Formally the National Climatic Data Centre
• https://www.ncdc.noaa.gov/isd
• https://gis.ncdc.noaa.gov/maps/ncei/
The Integrated Surface Database, which is publically available via the National Centres
for Environmental Information, consists of global hourly and synoptic observations
compiled from numerous sources into a single common ASCII format and common data
model. ISD integrates data from over 100 original data sources.
The database includes over 35,000 stations worldwide, with some having data as far
back as 1901. Currently, there are over 14,000 active stations updated daily in the
database.
ISD includes numerous parameters such as wind speed and direction, wind gust,
temperature, dew point and cloud data for various time periods and for various other
elements as observed by each station.

What about other Global Resources?
• The International Renewable Energy Agency (IRENA) Global Atlas for Renewable
Energy http://irena.masdar.ac.ae/#
• Renewables.Ninja https://www.renewables.ninja/#
Two other global resources are the International Renewable Energy Agency’s (IRENA)
Global Atlas for Renewable Energy and the Renewables.Ninja calculator.
IRENA’s Global Atlas for Renewable Energy provides access to a variety of different maps
of the solar and wind resource. For example, for solar, annual averaged maps of GHI are
available from the BoM, Meteotest and Vaisala.

The World Bank Global Solar and Wind
Atlases.
• http://globalsolaratlas.info/
• https://globalwindatlas.info/
A third global resource is the World Bank Group’s Global Solar Atlas which maps annual
values of GHI, DNI, DHI (DIF) and POA irradiance (GTI) as calculated from SolarGIS’
satellite derived irradiance data product. The platform was updated in late 2019 an now
allows users access to DNI monthly averages and daily average profiles. This map can be
accessed for free through the World Bank Group’s Global Solar Atlas
http://globalsolaratlas.info/
And for Wind there is the World Bank Group’s Global Wind Atlas which maps wind
speed, wind rose’s and power density at heights of 10, 50, 100, 150 and 200m. This map
can be accessed for free through the World Bank Group’s Global Wind Atlas
https://globalwindatlas.info/

Activity 1: Weather Data Sources
Let’s explore a few of the weather data sources we’ve just learnt about by completing
our first in-class activity.

Activity: Weather Data Sources
1. Pair up with a fellow classmate.
• Select any location within Australia and define that location by its latitude and longitude.
• Ensure that you and your fellow classmate/s choose locations in different parts of the country.
• NB: You could use google maps to choose a random location and identify its latitude and longitude
https://maps.google.com/
2. Do: For your chosen location, calculate the annual daily average of Global Horizontal Irradiation
(GHI) in kWh/m2.day and Air Temperature in degrees Celsius using the following online resources:
• The BoM’s Climate Data Online service http://www.bom.gov.au/climate/data/
• NASA’s POWER Data Access Viewer https://power.larc.nasa.gov/data-access-viewer/
• The World Bank Group’s Global Solar Atlas http://globalsolaratlas.info/

Activity: Pair, Do, Think, Re-Pair, Share
3. Think about the results you’ve calculated. Consider the following questions:
• How did you go about calculating the average air temperature when using the BoM Climate Data
Online?
• How well do the three data sources agree with one another?
• Are the results what you expected with regards to both the magnitude (i.e. average) and variation
(i.e. range, standard deviation)? If not, why not?
4. Re-Pair with your fellow classmate and Share and Discuss the results you both observed. Again
consider the above three questions but this time also consider:
• How do your results differ from your classmates?
• Are these the results you expected? If not, why not?

Example - UNSW How important are these observed differences?
GHI (kWh/m2.day) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual
BoM (1990-2020) 6.50 5.70 4.90 3.90 2.90 2.30 2.70 3.70 4.80 5.60 6.20 6.60 4.60
NASA (1983-2005) 5.94 5.28 4.55 3.59 2.79 2.52 2.72 3.61 4.65 5.54 5.91 6.28 4.45
SolarGIS (2007-2018) 4.54
Average 6.22 5.49 4.73 3.75 2.85 2.41 2.71 3.66 4.73 5.57 6.06 6.44 4.53
Range 0.56 0.42 0.35 0.31 0.11 0.22 0.02 0.09 0.15 0.06 0.29 0.32 0.15
Range (%) 9.0% 7.7% 7.4% 8.3% 3.9% 9.1% 0.7% 2.5% 3.2% 1.1% 4.8% 5.0% 3.3%
Temperature (°C) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Annual
BoM (1939-2020) 22.9 22.8 21.5 18.7 15.6 13.2 12.3 13.3 15.7 18.0 19.9 21.8 18.0
NASA (1984-2013) 22.6 22.4 21.0 18.8 16.1 13.5 12.5 13.3 15.6 17.7 19.4 21.3 17.8
SolarGIS (2007-2018) 17.7
Average 22.7 22.6 21.3 18.7 15.8 13.3 12.4 13.3 15.6 17.8 19.6 21.5 17.8
Range 0.26 0.45 0.47 0.17 0.46 0.38 0.27 0.03 0.08 0.31 0.47 0.50 0.25
Range (%) 1.1% 2.0% 2.2% 0.9% 2.9% 2.8% 2.2% 0.2% 0.5% 1.7% 2.4% 2.3% 1.4%
For the example presented on this slide I’ve used the location of UNSW’s Village Green with
coordinates of -33.918369, 151.227937.
Using just three data sources, this simple analysis indicates that the annual values from all three
sources differ from one another.
For GHI, the magnitude of difference (i.e. the range) of the annual figures is 0.2 kWh/m2.day,
which is equivalent to 3.3% of the mean of the three data sources. For air temperature, the
difference in the three annual figures is 0.35C, or 1.4% of the annual value.
When we look at the monthly values we can see even larger differences. For example, for the
month of January the difference in the GHI values is 9% of the mean value. Similarly, for air
temperature we observe the largest difference occurs for the month of May, equal to 2.9% of
the mean value.
These results raise some important questions, for example:

• What’s causing these differences?
• Which value is the true value?
• Which value do we use in our modelling?
• Should we be using just one or multiple data sources?
Perhaps you had other similar questions. We will explore these questions in future lectures. At
this point you simply need to be aware that numerous weather data sources exist and that their
values may not necessarily agree with each other.

7
6
Average Daily Irradiation (kWh/m2.day)
4
BoM (1990-2020)
3 NASA (1983-2005)
SolarGIS (2007-2016)
2
0
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Month
Sydney
This slide presents the monthly results of daily average irradiation (or insolation) from
our in-class activity as a simple line graph.
NB: Although the monthly SolarGIS data is not publicly available, they are available for
purchase. I’ve included the monthly values for our example location on this graph as a
third comparison.
When we look at the data visually on a graph, we can gain a greater perspective of the
magnitude of difference between the three data sources.
Depending on what weather data source we choose, this difference in the observed
daily level of irradiation would have a significant impact on any energy modelling we
undertake.

24
22
Average Daily Temperature (C)
20
18
BoM (1939-2020)
16
NASA (1984-2013)
14
12
10
Month
Sydney
This second graph presents the daily average values for air temperature. Again by
presenting the data visually as a graph, it quickly becomes apparent that the largest
difference in the two data sources occurs for the months of May and June.

Analysis: Visual Comparison of Annual Global Horizontal Irradiation
Our previous slides, looked at three data sources (the BoM, NASA and SolarGIS) for the
single location of UNSW, Sydney. Now, to further explore the concept that multiple data
sources can disagree, this slide presents maps of annual average global horizontal
irradiation across Australia, in units of kWh/m2.year, from six different data sources.
What do you observe or conclude from looking at these maps?
Each of these maps present the long-term annual average value of global horizontal
irradiation, but as was evident in our point based example, the data sources don’t agree
with one another, and the level of agreement or disagreement (and hence uncertainty)
is dependent on the location of interest.
Did you also happen to observe that the averages were calculate across different
temporal ranges? For example, the map of GHI from the BoM is the average over the
period 1990 to 2017 (both inclusive), whilst the map from Meteonorm 7.2 is the average
taken across the period 1991 to 2010.
Calculating annual averages over different historical periods can explain part of the
observed differences across these data sets, but it doesn’t account for all of the
differences observed. There are several other factors at play in this example, which we
will learn more about in future topics, including the spatial resolution of the data and
the accuracy of the underlying method used to generate the data.

Relative Standard
Mean
Deviation
Analysis: Mean and Relative Standard Deviation of Annual GHI

Data Sources: BoM, Meteonorm 7.2, NASA POWER, Solargis and Vaisala (MERRA-2 excluded)
Continuing on with our example, this slide presents the average and relative standard
deviation of five of the maps that were presented on the previous slide (MERRA-2
excluded).
The map of relative standard deviation illustrates that the magnitude of the relative
standard deviation, and hence the level of uncertainty across these five data sets, is not
constant across Australia. This example further illustrates how multiple data sources can
disagree, raising questions about the accuracy and uncertainty of the various data
sources, whilst also raising questions about which data source to use and when.
We will explore the topic of uncertainty and revisit this example when we explore the
topics of resource assessment and bankable data sets. For now, the key message is to
understand that several data sources exist and to begin to understand why it is
important to assess weather data for renewable energy modelling.
NB: MERRA2 was excluded from the analysis presented on this slide, as it is an outlier,
when viewed in comparison to the other five data sets which were presented on the
previous slide.

Weather Files

Weather Files – TMY
• Provides a single year of hourly data
• Originally developed for solar heating and cooling applications i.e. building modelling
• Based on historical data
• Typically lacks current and/or recent real-time meteorological and solar information
• Lacks information about climate variability of the particular location
• Does not include the influence of extreme events
Example: NSW Dust storm 2pm on the

23rd Sept 2009 significantly reduced the
level of irradiance.
Weather data from national meteorological providers is not the only source of weather
data available. Historically, the most common form of weather data used in RE system
modelling, particularly for building energy and PV system modelling, are pre-fabricated
weather files, the most common of which is the Typical Meteorological Year or TMY
weather file.

• Numerous TMY weather data files available:
– IWEC/IWEC2 – International Weather for Energy Calculations
» IWEC2 - 3012 files worldwide; 103 for Australia
– AREMI TMY – Australia (Only 6 locations available)
– RMY – Australia Representative Meteorological Years
– ERMY – Exemplary Representative Meteorological Years for Australia
– NatHERS RMY Climates files
– Similar products available for other countries (TMY/2/3, CWEC, NIWA, CSWD etc.)
• Access to free TMY weather files via:

– U.S. DOE EnergyPlus website https://energyplus.net/weather
– Climate.OneBuilding.Org http://climate.onebuilding.org/default.html
Various TMY weather files have been developed over the past couple of decades. Each
series of weather files were developed using different weather databases, different
historical periods of record and different weightings of the weather parameters.
Some examples of TMY weather files available globally include:
The IWEC and IWEC2 weather files, where the most recent version, the IWEC2 files,
were developed from 23 years of data from the Integrated Surface database. It should
be noted that the solar radiation in these weather files were estimated on an hourly
basis from measurements of cloud cover.
Some examples of TMY weather files available in Australia include:
The weather files generated by AREMI using the BoM’s One Minute Solar irradiance and
AWS data, the set of Australian RMY weather files developed for the Australian
Greenhouse Office for use in complying with the Building Code of Australia, the 2012
NatHERS reference meteorological year climate files compiled from BoM climate data
and the ERMY developed by Exemplary Energy utilising 23 years of satellite derived
irradiance from the Australian BOM.
The original version of the IWEC and RMY weather files are available free to download
from the US. Department of Energy’s EnergyPlus website and a copy of the IWEC2
weather files are also available through UNSW’s library.

• Provides a single year of hourly data
• Originally developed for solar heating and cooling applications i.e. building modelling
• Based on historical data
• Typically lacks current and/or recent real-time meteorological and solar information
• Lacks information about climate variability of the particular location
• Does not include the influence of extreme events
Example: The dust storm which hit NSW in September

2009 significantly reduced the level of irradiance.
This event would exclude September 2009 from being

included within a TMY weather file.
So what is a Typical Meteorological Year Weather file?

Generally, a TMY weather file contains a single year worth of hourly weather data i.e.
8760 data points for each weather variable. The data is usually selected from a historical
period of record, with the aim of reflecting the long-term average for the location in
question.
A TMY is NOT a single year of continuous weather data, such a weather file is known as
a Typical Reference Year. Rather a TMY is a single year of data where each month was
individually selected to reflect that months long term average.
For example, a TMY weather file may contain the weather data from 1990 for the month
of January, the data from 2004 for the month of February, and the data from 1997 for
the month or March and so on. The process of creating a TMY weather file is quite
intensive and requires a detailed statistical analysis. We will briefly take a look at this
process later in this lecture.
It should be noted that TMY weather files usually lack current or recent real-time data,
since they are created at a fixed point in time and are rarely updated. Similarly, because
these weather files only contain a single year of data, they also lack information about
climate variability and usually ignore the influence of extreme events such as volcanic
eruption or climatic phenomenon that occur at decadal or longer timescales, for
example the effects of El Nino and La Nina in Australia. TMY’s would also exclude other
extreme events like dust storms as per the dust storm that occurred across NSW and
Victoria in September 2009.
It should also be noted that not all of the meteorological parameters contained within a
weather file are measured. Usually, the irradiance parameters are modelled based on
cloud cover measurement or are derived from satellite imagery.

Activity 2 – Explore a TMY Weather File
Let’s explore a TMY weather file for our second in-class activity.

Explore a TMY Weather File
• Download the Australian RMY weather file for the location of
Sydney Airport
https://energyplus.net/weather-location/southwest_pacific_wmo_region_5/AUS//AUS_NSW.Mascot-
Sydney.Airport.947670_RMY
– How many rows of hourly data does the file contain?

– What was the most recent year of data included in the file?
– When was the weather file created?
– What period of data was considered in the files creation?
Let’s take a look at an example weather file for the location of Sydney Airport. See if you
can answer the questions listed on this slide.
You want to download the “epw” file. To view this file in a text editor or Excel, simply
change the file extension from “.epw” to “.csv”. NB: Whilst TMY refers to a type of
weather file “epw” refers to how the data is formatted within the file. “epw” is just one
of many weather file formats.
Let’s first confirm that the file provides a single year of hourly data. We can easily do this
by looking at the total number of rows of data within the file. In this file there is a total
of 8768 rows of data. This consists of 8 header rows and 8760 rows of hourly data. We
can also tell the data is in an hourly format by looking at the “Hour” (3rd column) and
“Minute” (4th column) fields of the data in rows 9 to 8768.
Next, we can scroll through the hourly data or use the Max function in a program like
Excel to determine the most recent year of data included within the weather file. You
should find that the most recent year of data contained within this weather file was
1994!
In this particular weather file we can determine when the file was created and which
period of data was used in its creation by looking at the “Comments 1” field within the
header row (row 6). This particular weather file indicate that is was created in December
2006 using data from 1967-2004 i.e. the last 13 years of data was not considered.
The “Year” field can often be used to determine if a weather file is a TMY. For illustrative
purposes I’ve added some column headers to the snapshot of the data presented on this
slide. The first two columns correspond to the Year and Month of the hourly data. In this
case, we can see that the data from 1986 was selected for the month of January whilst
the data from 1994 was selected for the month of February, indicating that this weather
file is NOT a single year of continuous weather data.

7
6.5
Average Daily Insolation (kWh/m2.day)
5.5
4.5
BOM - 947670
4 NASA
3.5 SolarGIS
IWEC - 947670
3
RMY - 947670
2.5 IWEC2 - 947670
2
Month
Sydney
Up to this point we’ve discovered a number of sources where weather data can be
obtained for locations in Australia. We already looked at the difference between the
BoM and the NASA data for the location of UNSW, let’s now expand that analysis to
include the TMY weather data available for Sydney. In this particular example I’ve used
the data for Sydney Airport.
This first graph compares the monthly averages of daily irradiation as reported by the
BoM, NASA, SolarGIS and as reported in the IWEC, IWEC2 and RMY weather files, for the
location of Sydney. From this example we can see that differences up to 0.9 kWh/m2.day
occurs across the different weather sources at the monthly level. Depending on which
weather data source we choose, this difference in the observed daily levels of irradiation
would have a significant impact on any energy modelling we undertake.
NB: If we exclude the IWEC2 weather file the maximum difference observed across the
remaining five datasets reduces to 0.66 kWh/m2.day, but this level of difference can still
have a large impact on your energy modelling.

7
6.5
5.5
4.5
BOM - 945780
4 NASA
3.5 SolarGIS
IWEC - 945780
3
RMY - 945780
2.5
IWEC2 - 945790
2
Month
Brisbane
A second example in Brisbane illustrates how even greater discrepancies can occur
between the long term monthly average statistics for daily irradiation.
In particular, the satellite derived solar exposure from the BoM and SolarGIS result in
higher levels of daily insolation than detailed by the other data sources. These observed
difference can be caused by a range issues.
For example, part of the difference between the BoM, NASA and SolarGIS data will be
caused by the spatial resolution of the gridded dataset, i.e. the higher resolution 5 x 5km
grid from the BoM versus the 100 x 100km grid of the NASA’s dataset. The difference
will also be contributed to based on the period of data used to calculate the long term
average, for example a 27 year dataset (1990 to 2016) for the BoM data versus a 22 year
dataset (1883 to 2005) for the NASA data, and a 10 year period for the SolarGIS data
(2007 to 2016).
Similarly, the accuracy of the underlying data will also impact the results, recall how
irradiance data in TMY files are usually a modelled value based on inputs of cloud cover
and how satellite derived irradiance is also a modelled dataset.
Again, the GHI data within the IWEC2 weather file appears to be a significant outlier in
comparison to the other available data sources.

24
22
Average Daily Temperature (°C)
20
18
16
BOM - 947670
NASA
14
IWEC - 947670
12 RMY - 947680
IWEC2 - 947670
10
Month
Sydney
The second set of examples compares the monthly averages of daily air temperature as
reported by the Australian Bureau of Meteorology, NASA and as reported in the IWEC
and RMY weather files.
In this case the data from the BoM is a measured dataset, where different periods of the
historical record were used in the creation of both the IWEC and RMY weather files.
Hence we see a reasonably good agreement between these three data sources as they
are all based on measured data sources. The monthly statistics from NASA are a
modelled gridded dataset on the 100x100km grid.

30
28
26
Average Daily Temperature (°C)
24
22
20
18 BOM - 945780
16 NASA
IWEC - 945780
14
RMY - 945780
12
IWEC2 - 945780
10
Month
Brisbane
Again the results from Brisbane confirm the disagreement observed between the
monthly average statistics reported from the NASA gridded dataset, in comparison to
BoM measured data.
These two examples, illustrate why you need to carefully consider and assess the
weather data you are going to use in your renewable energy modelling.
For example, if we were modelling a PV system using PVsyst, it wouldn’t be the best idea
to automatically import monthly average statistics from NASA SEE if our location of
interest was Brisbane.

8
7
Average Daily Wind Speed (m/s)
3
BOM - 947670 (Mean 9am)
BOM - 947670 (Mean 3pm)
2
NASA
IWEC - 947670
1 RMY - 947680
IWEC2 - 947670
0
Month
Sydney
Finally, this third set of examples compares the monthly averages of wind speed as
reported by the Australian Bureau of Meteorology at 9am and 3pm, by NASA at a height
of 10m and as reported in the IWEC and RMY weather files. This example demonstrates
how significant differences occur in the wind speed data.

8
7
Average Daily Wind Speed (m/s)
3
BOM - 945780 (Mean 9am)
2 BOM - 945780 (Mean 3pm)
NASA
IWEC - 945780
1
RMY - 945780
IWEC2 - 947670
0
Month
Brisbane
Similar levels of differences are again observed for Brisbane.
When it comes to building energy or PV system modelling, this level of observed

difference in the wind speed data is unlikely to cause any significant differences in your
system energy modelling. For example, in PV system modelling wind speed has a third
order effect on performance, after irradiance and then air temperature. It should also be
noted that a number of PV system models, like PVsyst, by default ignore the effect of
wind speed in the model. However, if you were modelling a wind energy system, then
your choice of wind speed data will have a significant impact on your results. It should
be noted that TMY and single year weather files are not typically used in wind energy
modelling, rather, long periods of historical data are utilised in conjunction with a
minimum of 1 year of onsite hub-height measurements of wind speed to create what is
known as a bankable data set. We will explore this topic in a future lecture. The practice
of recording a minimum of 1 year worth of onsite weather and combining it with
modelled historical data like the BoM satellite derived irradiance data set to create
bankable datasets, is also gaining traction in PV system modelling, particularly in the
development of large solar farms; however TMY weather files are still the dominant
source of weather data used in modelling residential and commercial sized PV systems.

Weather Files – TDY & TGY
• Typical Direct Year and Typical Global (horizontal) Year
– Designed primarily for system performance analysis of the solar components
– Difference between TMY, TDY and TGY are the weighting factors
– For TDY and TGY the weighting are only applied to the solar parameters.
– TMY includes non-zero weighting factors for both the solar and meteorological
components.
– TDY – 100% weighting for direct normal irradiance (DNI)
– TGY – 100% weighting for global horizontal irradiance (GHI)
– TDY more appropriate for concentrating systems, TGY more appropriate for standard
flat plate PV systems.
Note: There are no TDY and TGY weather files for Australia.
Two newer categories of TMY weather files are the Typical Direct Year and Typical Global
Year weather files.
These two sets of weather files are created in a similar manner as TMY weather files,
except the weightings of the meteorological variables used in the statistical process to
select each month, only considers direct normal irradiance for TDYs and global
horizontal irradiance for TGYs.
These categories of weather files were developed by NREL in the US for modelling
concentrating and tracking RE systems and for fixed tilt PV systems, respectively.
It should be noted that a set of TDY and TGY weather files does not currently exist for
Australia.
We will cover weightings and how to create a TMY weather file in more detail in a later
lecture.

TMY – Example Weightings
Parameter Sandia Method TMY2/TMY3 IWEC TGY TDY
Dry Bulb Temp Max 4.17% 5.00% 5.00%
Dry Bulb Temp Min 4.17% 5.00% 5.00%
Dry Bulb Temp Mean 8.33% 10.00% 30.00%
Dew Point Temp Max 4.17% 5.00% 2.50%
Dew Point Temp Min 4.17% 5.00% 2.50%
Dew Point Temp Mean 8.33% 10.00% 5.00%
Wind Speed Max 8.33% 5.00% 5.00%
Wind Speed Min 8.33% 5.00% 5.00%
GHI 50.00% 25.00% 40.00% 100%
DNI 25.00% 100%
For example, if we wanted to create a TGY weather file, the only parameter used to
identify the most appropriate month is the parameter of GHI. In this case the weighting
is 100% on the GHI component and 0% for all other meteorological variables.
This slide presents a sample of some of the weightings that have been used to develop
different sets of TMY weather files. The weightings for the TGY and TDY weather files
have been added for quick comparison.

Weather Files – TRY
• Test or Typical Reference Year
– 1 single continuous year of hourly measured data
– Where possible, avoid using TRY weather files
– No single year can represent the typical long term weather pattern.
Another type of weather file is the test or typical reference year.

This type of weather file contains a single year worth of continuous data, where the year
selected has the closest annual average to the long term annual average of the historical
record.
Generally, I do not recommended the use of a TRY weather file for long term
performance modelling, as any single year of data is unable to represent the average
long term monthly weather patterns.

8
Historical Average (1990-2019)
TMY (1991, 1999, 2010, 2013, 1991, 1994, 1991, 2004, 2015, 2004, 1992, 2005)
7
TRY (2009)
2
Month
Sydney
For example, lets take a look at the monthly historical averages of daily solar exposure
for Sydney airport. In this simple case, which only looks at the one weather variable of
GHI, if we apply the TRY method of selecting a single year of data which best matches
the long term annual average, we can end up observing significant differences at the
monthly level. Conversely, if we apply the TMY method and select each month
individually, we do a good job at matching both the long term annual and monthly
averages.

Weather Files - AMY
• Actual Meteorological Year
– Weather files developed for a specific year of weather data
– Usually user generated
– A number of commercial companies starting to create AMY weather files for specific
locations
» Exemplary Energy – Set of AMY weather files for Australia based on BOM AWS
data and satellite derived irradiance.
» SolarGIS with SolarGIS satellite derived irradiance
Actual meteorological year weather files, otherwise known as AMYs, are similar in
concept to TRY weather files, in that it is a single year of continuous weather data.
Unlike TRY weather files, AMYs are not meant to represent the long term average, but
rather represent the weather data for a specific year.
AMYs have many uses, from designing a system based on load data over a specified
period, to performance monitoring and analysis of RE systems, to the calculation of
probability of exceedance statistics through the use of multiple years worth of AMY
weather files.

Synthetic Weather Data
• The Case of PVsyst/Meteonorm:
– Synthetic data generation provides a means of constructing meteorological hourly
data from monthly inputs of GHI and temperature.
– Often used in the absence of a TMY weather file
– More on this topic in a future lecture…
Another type of weather data that is commonly employed in PV system modelling is

synthetically generated weather data.
This method makes use of monthly averages of insolation and air temperature to
generate a full year of hourly synthetic data. Both PVsyst and Meteonorm provide
functions to synthetically generate weather data. It should be noted that such a feature
is not currently available for modelling in NREL’s SAM.
We will cover synthetic weather data in more detail in a later lecture and you will also
learn how to use PVsyst to generate a synthetic weather file in a later tutorial activity.

What about other Countries?
• The United States:
– NREL has over time developed the National Solar
Radiation Database
» http://rredc.nrel.gov/solar/old_data/nsrdb/
» https://maps.nrel.gov/nsrdb-viewer/#
 Data for India is also available
• PVGIS for Europe and Africa:

– http://re.jrc.ec.europa.eu/pvgis/apps4/pvest.php
– http://re.jrc.ec.europa.eu/pvgis/apps4/pvest.php?lang=en
&map=africa
This module focused primarily on the location of Australia, with a few mentions of global
resources like NASA’s SSE, MERRA 2, ERA5 or the Integrated Surface Database. If
Australia is not your nation of interest, you might be asking what weather data
resources are available for my country.
Here are a few resources that I know of for the United States and Europe and Africa.

• United Kingdom:
– UK Met Office
» https://www.metoffice.gov.uk/
• Saudi Arabia:
– K.A.CARE created the Renewable Resource Atlas for
Saudi Arabia
» https://rratlas.kacare.gov.sa/RRMMPublicPortal/?q=en
• Indonesia:
– Indonesian Agency for Meteorology, Climatology and
Geophysics (BMKG) map the location of AWS data in
Indonesia.
» http://202.90.199.132/aws-new/
And here are a few resources that I know of for the UK, Saudi Arabia and Indonesia.

Do you know of any other resources?
• India:
– Open Government Data (OGD) Platform India
» https://data.gov.in/catalogs/ministry_department/india-meteorological-department-imd
• Chile:
– Direccion Meterologica de Chile
» www.meteochile.cl
– Servicio Metereologico de la Armada de Chile
» http://meteoarmada.directemar.cl/prontus_meteo/site/edic/base/port/inicio.html
– SINCA Sistema de Información Nacional de Calidad del Aire
» https://sinca.mma.gob.cl/
If you know of specific weather data sources available for your country please let the
course coordinator know. This will help improve the lecture content in future years.
For example, these resources for India and Chile have been added to this list of available
weather data sources based on contributions made from students in the 2018 class of
SOLA9103.

Paid By Demand Services
• Other paid by demand services for weather data are available from commercial
companies including:
– Meteonorm, SolarGIS, 3Tier/Vaisala, AWS Truepower, Green Power Labs, Solcast
Finally, numerous consulting companies exist which offer paid services to create TMY
(Typical Meteorological Year) and AMY (Actual Meteorological Year) weather files or
provide long term historical databases of irradiance, temperature and wind speed.
It should be noted that in many cases the weather variables available from these
commercial companies will consist of modelled or satellite derived irradiance data, and
air temperature and wind speed data which may be interpolated for your specific
location from nearby ground measurement data.

Summary
By now you should be able to:
• Identify a variety of available weather data sources and associated issues
• Describe the difference between TMY, TRY, AMY, TDY & TGY weather files
• Identify why it is important to assess weather data for RE system modelling.
You should now be able to identify a variety of weather data sources; describe the
difference between various types of weather files; and begin to identify why it is
important to assess weather data for RE system modelling. In future lectures we will
explore the differences in these data sources in more detail we when explore the topics
of Resource Assessments and Bankable Datasets.


Topic 1.1 Weather Data Annotated PDF

Uploaded by

Copyright:

Available Formats

Topic 1.1 Weather Data Annotated PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Topic 1.1 Weather Data Annotated PDF

Uploaded by

Copyright:

Available Formats

Dr Jessie Copper

School of Photovoltaic and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

Topic 1.1: Weather Data Sources

The first topic in SOLA9103 is on weather data sources.

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

What is the long-term average value of global

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

How Would You Respond?

4.6 4.45 4.59 4.54 4.51 5.2 ERMY – 4.62

School of Photovoltaics and Renewable Energy Engineering

So, how did you respond to the question?

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

A Temporal and Point location dataset

School of Photovoltaics and Renewable Energy Engineering

Let’s now take a look at some specific sources of data.

SOLA9103 – RE System Modelling and Analysis

A Temporal and Point location dataset

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

A Temporal and Point location dataset

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

– Maps – Recent conditions http://www.bom.gov.au/climate/maps/

A Geospatial or Gridded dataset

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

The Australian Renewable Energy Mapping Infrastructure, otherwise known as AREMI,

SOLA9103 – RE System Modelling and Analysis

SOLA9103 – RE System Modelling and Analysis

SOLA9103 – RE System Modelling and Analysis

SOLA9103 – RE System Modelling and Analysis

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

So why then do I highlight the NASA dataset?

SOLA9103 – RE System Modelling and Analysis

• Other reanalysis datasets available worldwide include:

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

A Geospatial or Gridded dataset

School of Photovoltaics and Renewable Energy Engineering

What is climate reanalysis? by ECMWF ‘A climate reanalysis gives a numerical

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering

SOLA9103 – RE System Modelling and Analysis

School of Photovoltaics and Renewable Energy Engineering