[NOTE: This project is no longer available, and has been undertaken by Sonya Wellby of the Fenner School of Environment and Society]
The meteorological events that impact solar energy generation are my area of expertise, so naturally I find them intensely interesting. So why not a blog-post on the topic? Let's do it!
There are many different types of cloud events that impact solar photovoltaic electricity generation it a wide variety of ways. From thick stratus decks to thin high cirrus – there are plenty of things to research! And today is your lucky day, because I want to share with you one of the types that I am most interested in: fog/low-cloud events
Why do we care?
So why the interest? Well, first, fog events are very common in Canberra, leaving the city shrouded in thick low clouds with limited visibility many of the mornings throughout the year. Just how common? Let’s take a look at all a distribution of the relative humidity and wet-bulb measurements over the past 3 years as observed at the Canberra airport:
I’d say that’s a pretty convincing set of histograms – Canberra certainly sees a large number of periods with high humidity/near saturation conditions. These are quite often (but not always!) associated with fog.
This type of low cloud/fog hangs around until enough diffuse solar radiation has passed through it to re-start the boundary layer convective mixing process. Once this gets going, the fog/low level cloud decks tend to ‘burn off’ very quickly. How long it takes for this process to get started depends on how thick the low-level cloud deck is and whether or not there are also upper/mid level cloud decks in place to further reflect/scatter the solar radiation traversing the atmosphere. Here is an example of a quick burn-off event, taken from a rooftop PV site:
Here, the red-portion of the time series is meant to represent when fog occurred [RH>85,WSP<10,WTB_DEP<3,CLEAR_SKY=FALSE], the blue is the power output from the PV array and the green is the relative humidity. We can see a sudden and sharp increase in power output with a sharp drop in relative humidity. Very cool!
I’m also quite keen to better understand these types of events, since they happen in the early morning, coinciding with the first peak in home energy demand. Just as everyone is waking up, starting their tea kettles/coffee pots, flipping on the heaters and lights and getting their day started, our rooftop solar generation is stuck in low gear. Rooftop solar arrays tend to operate at less than 40% of their clear sky potential under thick fog – and thus this type of weather event is very problematic for solar energy generation. Here is a plot for the same day with that clear sky potential (KPV) plotted:
In addition to the overall reduction in solar energy generation that occurs during this morning demand period, there is often a corresponding sudden ramp event that occurs as these clouds clear (like we just discussed). The mixing process that drives the convective overturning of this stable cloud layer is a positive feedback process. As it starts, more cloud is mixed out, so more radiation makes it through, the ground heats more, convection intensifies and more cloud is mixed out – and so on and so forth. The result is a very quick disappearance of the low level cloud/fog. This type of ramp event could be troublesome for areas with high PV penetration – as solar generation sites all suddenly produce more power. In fact, I have some photos from the Canberra Yacht Club webcam that show this on 12 September 2012, let's take a look:
The cloud is very thick, but completely disappears within an hour - even with the presence of an upper-level cloud deck.
Furthermore, this convective mixing process happens unevenly throughout the region. In Canberra, there is a lot of variation in topography, several lakes and a large geographic spread in the suburban area. So correctly capturing where and when the cloud/fog will first mix out is a challenging problem. You can see this from the ACT Solar Map graphics at 9:20AM and 10:10AM:
Notice how the fog/low-cloud dissipates unevenly? This is also somewhat evident in the satellite imagery.
For this reason, I am currently recruiting an Honours Student from within the Fenner School of Environment and Society to undertake a project in this area. The chosen candidate will analyze both weather and photovoltaic power data from the Canberra region to:
1) Identify low-cloud/fog events
2) Establish the origin and meteorological conditions leading to their formation
3) Seek to describe the behaviour of the transition period that occurs as low-cloud/fog burn off
4) Assess the impact of these events on collective PV power production
The skills gained in taking on this project will be invaluable for any student interested in pursuing a career/further studies in meteorology and or energy related fields.
If you’re interested, send me an email!