How to consider weather condition for daylight simulation in Radiance

Hi everyone,

I recently learned to use Radiance for interior daylight simulation. I read the first 6 chapters of Rendering with Radiance (1998), but there is no clear introduction to consider weather conditions for daylight simulation.

I have read the topic related to the weather before (weather data in radiance - #3 by Lars_Grobe). But I am still confused with the following aspects:

  1. Can I assume that when the gensky is used to generate the sun and the sky, it already considers the weather conditions? Because the irradiance database that gensky used to define the sky and sun, already considers weather conditions (the equipment to record the irradiance is on the ground so the data it recorded considers the influence of different weather)

  2. For daylight simulation, should we distinguish the weather conditions and atmosphere conditions.

  3. My understanding of GENDAYLIT is that it mainly considers different atmosphere conditions which my already considered when the sun data was recorded.

Thank you in advance for your help.

Best regards,
Hao

Hi @Hao,

When we talk about “weather” in the context of Radiance simulations, we are usually talking about the level of cloudiness (absorption and diffusion) in the atmosphere. We do not consider other factors like rain. So for our purposes, weather and atmosphere are the same.

The GENSKY program produces clear, overcast, and intermediate skies according the their CIE definitions. This is useful if you want to render a space under sunny conditions (e.g., for shadow visualization or glare simulation) or overcast conditions (e.g., for daylight factor calculation).

The GENDAYLIT program produces skies according to the Perez All-Weather Sky Model. This is useful if you want to match the recorded direct and diffuse components from TMY weather data files.

3 Likes

Hey Hao,

On point 3, you mentioned atmospheric conditions. gendaylit does not directly model atmospheric parameters such as precipitable water vapor or aerosols. Instead, it focuses on generating sky, sun, and ground irradiance (or illuminance) based on the provided direct and diffuse irradiance components (DNI and DHI). These irradiance components inherently account for the effects of atmospheric conditions, such as cloud cover, scattering, and absorption.

Additionally, it’s important to distinguish between Typical Meteorological Year (TMY) data and weather data. TMY provides a representative year of meteorological data. In contrast, weather refers to actual measurements of meteorological parameters taken during specific periods.

Both will be accepted by gendaylit and the choice depends on the purpose of your study.

Cheers,
Alex

Note:
If you want to have better control over atmospheric conditions you can check genssky. However, I have not yet experimented with it.

2 Likes

Hi @Nathaniel_Jones and @Funerus,

Thank you so much for your explanation which gives me a clearer understanding of the function of Radiance. I will learn the gendaylit and genssky programs and the TMY data. I truly appreciate it.

However, I am still confused about the following point: when the data of the sun was recorded, the data already includes the atmospheric condition (the equipment to record the irradiance is on the ground ). So when the gensky command uses this sun data to generate the sun and sky, it already considers the atmospheric conditions. So why should consider the atmospheric condition again? I don’t have much experience in daylight analysis, so this question might seem foolish.

Or the solar data is from the satellite, so when the gensky is used to generate sun and sky, with defaults (its default is clear CIE sky, ), the radiance data of solar is the data in the perfect atmospheric condition(without cloud); for the sky distribution, the -t 2.450000 # Atmospheric turbidity is used to give sky daylight distribution and determines the diffuse horizontal irradiance.

Best regards,
Hao

Hi @Hao,

The measurement equipment is located at ground level. However, the measurement equipment only records global horizontal irradiance and direct normal irradiance (or, more likely, it records other data and uses algorithms to calculate GHI and DNI). However, GHI and DNI do not provide enough information to generate the sky distribution that Radiance needs. Instead, you must provide additional information to describe how the diffuse irradiance is distributed across the sky dome.

If you use GENSKY, then you will specify a CIE sky type that determines the distribution, and you may optionally specify turbidity directly to modify the distribution away from the standard. If you use GENDAYLIT, then the distribution is produced according to the Perez model based on the direct and diffuse components and solar angle.

2 Likes

A minor correction – the turbidity in gensky does not affect the sky distribution, only the calculated zenith brightness. The CIE models are designed to be simple for reproducibility that way.

Therefore, there is no need to specify the turbidity if you know the diffuse irradiance or zenith radiance, as those specifications override the default calculation. This is in fact the recommended way to use gensky if you have measurements.

1 Like

Hi @Nathaniel_Jones ,

Thank you very much for your reply.

Your explanation is very helpful. I will learn the corresponding concepts and background knowledge to gain a deep understanding of how to deal with weather conditions.

Best regards,
Hao

Hi @Greg_Ward,

Thank you so much for your correction. That is very helpful.

Currently, I don’t have any measurements. I plan to use TMY data as the weather condition. Then set a day having 7 different weather conditions. For each weather condition, simulate the interior lighting condition of a room. In this case, could you please give me some suggestions for the program I can use? Or the GENDAYLIT and the Gendaylit are good choices?

Best regards,
Hao

I would definitely recommend gendaylit over gensky if you are attempting to simulate different weather conditions, because that’s what gendaylit’s Perez sky model is designed to do. The CIE sky model is more intended to create reproducible comparisons between designs under “typical” skies, and gensky does not contain any of the more advanced CIE intermediate sky types.

Cheers,
-Greg

1 Like

Hi Greg,

Thanks a lot for your recommendation and helpful explanation. I will explore gendaylit and the Perez sky model as you suggested.

By the way, for the standard to introduce the CIE sky models, is the standard BS ISO 15469: Spatial distribution of daylight — CIE standard general sky?

Cheers,
Hao