Milestone 2 - Radiation modeling

Introduction

The energy balance of Earth is strongly impacted by radiation. The following figure shows the energy fluxes in the global Earth atmosphere system.

  • Earth's Energy Budget. Source: NASA, Public domain, quoting Loeb et al., J. Clim 2009 & Trenberth et al, BAMS 2009.

Remark: Note that this figure is from 2010. By now the net absorbed energy has risen to 1.0 W/m21.0\; W/m^2. (Loeb et al. (2021). Satellite and Ocean Data Reveal Marked Increase in Earth’s Heating Rate. Geophysical Research Letters, 48(13))

In our EBM, we aim to include four effects:

(i) Incoming solar radiation.

(ii) Reflection of radiation by the surface and cloud cover.

(iii) Cooling because of outgoing longwave radiation.

(iv) Effect of greenhouse gases (mainly CO2CO_2).

Outgoing longwave radiation

The goal in this section is to define a model/parametrization for the source term SOLW(T,x,t)S_{OLW}(T,x,t).

But first, to warm up with the topic, we consider as a toy model an idealized black-body Earth, i.e., we assume that the Earth is a black body that emits and absorbs radiation in the infrared spectrum.

The radiation energy per time that is emitted by a black body can be computed by the Stefan-Boltzmann law of physics

I=σSB TR4, I = \sigma_{SB}\,T_R^4,

where II is the radiation energy per time per area with units [W/m2][W/m^2], TRT_R is the radiation temperature with physical units Kelvin [K][K], σSB=0.56687108\sigma_{SB} = 0.56687\cdot 10^{-8} is the Stefan-Boltzmann constant with units [W/m2/K4][W/m^2/K^4].

For this idealized black-body Earth, the temperature is determined when the outgoing radiation is in equilibrium/balance with the incoming stellar radiation. The amount of incoming energy can be roughly estimated as S0 (1α) π RE2S_0\,(1-\alpha)\,\pi\,R_E^2, where S0=1360[W/m2]S_0 = 1360 [W/m^2] is the solar constant (the mean solar elecromagnetic radiation received on Earth), α\alpha is the surface albedo (the amount of the solar radiation that is reflected back to space) with the planetary average being about α=0.3\alpha=0.3 (more details in the Albedo section), and RE=6.378106[m]R_E=6.378\cdot 10^{6} [m] is the radius of Earth. Note that the solar constant is scaled with the effective area in which the solar radiation is applied on earth: π RE2\pi\,R_E^2.

Taking into account that the surface of the sphere is 4πRE24\pi R_E^2, the amount of outgoing energy is σSBTR44 π RE2\sigma_{SB} T_R^4 4\,\pi\,R_E^2, and we get our first (simplest version of an) EBM

σSBTR44 π RE2=S0 (1α) π RE2. \sigma_{SB} T_R^4 4\,\pi\,R_E^2 = S_0\,(1-\alpha)\,\pi\,R_E^2.

We can relate this simple EBM (2) to the general EBM form by making the assumption of thermodynamic equilibrium (no temporal change T/t=0\partial T / \partial t= 0), no heat diffusion (d=0d = 0), and the following choice of source terms:

Ssol=S0 (1α) π RE2,SOLW=σSBTR44 π RE2.\begin{aligned} S_{sol} &= S_0\,(1-\alpha)\,\pi\,R_E^2, \\ S_{OLW} &= - \sigma_{SB} T_R^4 4\,\pi\,R_E^2. \end{aligned}

We can directly solve (2) to get the black-body equilibrium temperature

TR=(S04(1α)σSB)14255 [K]=18 [C]. T_R = \left(\frac{S_0}{4}\frac{(1-\alpha)}{\sigma_{SB}}\right)^{\frac{1}{4}}\approx 255\, [K] = -18\, [^\circ C].

This model is indeed as simple as it gets and it is no surprise that the quality of the prediction of an average Earth temperature is quite off. The black-body radiation temperature of Earth would be only about 18-18 degree Celsius, hence, some major modeling improvements are necessary.

Budyko's Empirical infrared Model

Budyko (1968) suggested an empirical linear model for the outgoing longwave radiation

IA+B T. I \sim A + B\, T.

It is in general motivated by available observational data, shown in the next figure

The figure shows infrared radiation density plots averaged monthly, measured by satellite compared to the surface temperature at the same month and location. (a) shows the whole sky (including clouds) and (b) shows only the clear (cloudless) sky.

If we consider the temperature in units Kelvin, we can fit the observed data with the linear model by choosing good constants AA and BB to get

IIR/OLW=A+B (T273.15), I_{IR/OLW} = A + B\,(T - 273.15),

with A=210.3A=210.3 as the radiative cooling in units [W/m2][W/m^2], and B=2.15B=2.15 the radiative cooling feedback with units [W/m2/K][W/m^2/K]. It is important to note that the choice of these parameters has a direct impact on the outgoing radiation and hence on the cooling. Several others have fitted the data differently, hence some range of choices for AA and BB is available. The values we select are from the paper by Zhuang et al. (2017).

We are now able to consider a second, but hopefully improved toy EBM. We replace the crude black-body radiation with a phenomenological approximation of the outgoing radiation (the Budyko model) to get

(A+B (TR273.15))4 π RE2=S0 (1α) π RE2, (A + B\,(T_R - 273.15)) 4\,\pi\,R_E^2 = S_0\,(1-\alpha)\,\pi\,R_E^2,

which can be directly solved again to get the equilibrium temperature

TR=S04(1α)AB+273.1513 [C]. T_R = \frac{\frac{S_0}{4}(1 - \alpha) - A}{B} + 273.15 \approx 13\,[^\circ C].
Remark 8: As discussed, we have introduced a data driven model for our outgoing longwave radiation energy change. We introduced an ansatz with parameters (parametrization) and used real world measurements/observations to select the parameters, i.e., to fit the model.
Remark 9: The phenomenological model of Budyko drastically improves the temperature prediction of Earth from the crude black-body temperature of 18-18 to the temperature 13C13^\circ C, compared to the observed value of about 14.5C14.5^\circ C.

Effect of greenhouse gases (CO2CO_2)

For the interested reader, we refer to chapter 4 of the book by Kim and North, "Energy Balance Climate Model", (2017, Wiley) and the following paper

Myhre, G., Highwood, E. J., Shine, K. P., & Stordal, F. (1998). New estimates of radiative forcing due to well mixed greenhouse gases. Geophysical research letters, 25(14), 2715-2718.

The Earth's atmosphere is composed of several gases, with the most abundant ones being nitrogen (N2N_2), oxygen (O2O_2), and argon (ArAr). These gases do not strongly absorb infrared radiation, which is important because the Earth's surface emits infrared radiation as it cools down after being heated by the sun. N2N_2 and O2O_2 are diatomic molecules, meaning they consist of two atoms chemically bonded together. Diatomic molecules have no permanent dipole, which means they have no separation of electric charge and therefore do not strongly interact with infrared radiation. Argon, on the other hand, is a monatomic gas, meaning it consists of individual atoms rather than molecules. However, it does not absorb infrared radiation because it does not have any modes of rotation or vibration in the infrared spectrum.

Overall, the lack of strong absorption of infrared radiation by these main constituents of the atmosphere allows heat to escape from the Earth's surface and be radiated out to space, helping to regulate the planet's temperature.

H2OH_2O molecules, on the other hand, have a permanent dipole moment because the distribution of electrons in the molecule is asymmetric. As a result, H2OH_2O molecules respond strongly to passing electromagnetic waves, including those in the infrared spectrum, which leads to the absorption of infrared radiation.

Molecules such as CO2CO_2, CH4CH_4, and NO2NO_2 do not have a permanent dipole moment because they are symmetrical in shape. However, they can still absorb infrared radiation through the phenomenon of induced dipole moments. When an infrared photon passes near one of these molecules, it can cause the electrons in the molecule to shift slightly, resulting in a temporary dipole moment. This temporary dipole moment can then interact with the passing electromagnetic wave, leading to the absorption of infrared radiation.

The ability of these molecules to absorb infrared radiation is significant because it allows them to contribute to the greenhouse effect. The greenhouse effect is the process by which certain gases in the Earth's atmosphere, including CO2CO_2, CH4CH_4, and NO2NO_2, trap heat and warm the planet's surface. Without this natural process, the Earth's average temperature would be much lower and life as we know it would not be possible. However, when these gases are present in excess, they can cause an imbalance in the greenhouse effect, leading to climate change.

Comment: The effect of greenhouse gases is very complex and there is a lot of ongoing research. H2OH_2O in particular is very important but ever so complex due to it being dependent on many other components and processes, i.e., the full water cycle. We will consider in our EBM only the effect of CO2CO_2 on the radiation. Infrared absorbtion means that the efficiency of our outgoing radiation (cooling effect) decreases (which causes a higher equilibrium temperature). This decrease in efficiency is topic of many investigations and studies, e.g., by the IPCC.

In this course, we follow the paper by Myhre et al. (1998) to define our parametrization of the greenhouse gas effect. From this paper, we first look at the effect of the amount of greenhouse gases, as plotted in the following figures:

  • Figures were generated with data from Myhre et al. (1998).

The figures show radiative forcing as a function of concentration in [ppmv][ppmv] ("parts per million by volume") for CO2CO_2 and in [ppbv][ppbv] ("parts per billion by volume") for CH4CH_4 and N2ON_2O.

The following table gathers simplified expressions to compute (fit) the radiative forcing caused by different greenhouse gases using data from Myhre et al. (1998) and the function

f(M,N)=0.47ln(1+2.01×105(M N)3/4+5.31×1015M(M N)1.52).\begin{aligned} f(M,N) = 0.47 \ln \left( 1 + 2.01 \times 10^{-5} (M \, N)^{3/4} + 5.31 \times 10^{-15} M (M \, N)^{1.52} \right). \end{aligned}
  • Table: Simplified expressions for the radiative forcing ΔA\Delta A in [W/m2][W/m^2] with coefficients of the IPCC report (1990) and Myhre et al. (1998). In the expressions, CC is CO2_2 concentration in [ppmv][ppmv], MM is CH4_4 concentration in [ppbv][ppbv], NN is N2_2O concentration in [ppbv][ppbv], and XX is Chlorofluorocarbons (CFCs) concentration in [ppbv][ppbv]. The subscript 00 denotes unperturbed (reference) concentrations. The function ff is given in (9). Adapted from Myhre et al. (1998).

Trace gasRadiative forcing ΔAαIPCCαMyhreCO2αln(C/C0)6.35.35CH4α(MM0)(f(M,N0)f(M0,N0))0.00360.0036N2Oα(NN0)(f(M0,N)f(M0,N0))0.140.12CFC-11α(XX0)0.220.25CFC-12α(XX0)0.280.33\begin{aligned} \begin{array}{cccc} \hline \text{Trace gas} & \text{Radiative forcing $\Delta A$} & \alpha_{\text{IPCC}} & \alpha_{\text{Myhre}} \\ \hline \text{CO}_2 & \alpha \ln (C/C_0) & 6.3 & 5.35 \\ \text{CH}_4 & \alpha \left(\sqrt{M}-\sqrt{M_0} \right) - \left(f\left(M,N_0\right)- f\left(M_0,N_0\right) \right) & 0.0036 & 0.0036 \\ \text{N}_2\text{O} & \alpha \left(\sqrt{N}-\sqrt{N_0} \right) - \left(f\left(M_0,N\right)- f\left(M_0 ,N_0\right) \right) &0.14 & 0.12 \\ \text{CFC-11} & \alpha(X-X_0) & 0.22 & 0.25\\ \text{CFC-12} & \alpha(X-X_0) & 0.28 & 0.33\\ \hline \end{array} \end{aligned}

As mentioned, we only consider the effect of CO2CO_2 in our model and hence choose the approximation

ΔA=αMyhre ln(CO2/CO2(t0)), \Delta A = \alpha_{\text{Myhre}}\, \ln(CO_2/CO_2(t_0)),

where

αMyhre=5.35 [W/m2], \alpha_{\text{Myhre}} = 5.35\,[W/m^2],

CO2CO_2 is the concentration in [ppm][ppm], and CO2(t0)=315 [ppm]CO_2(t_0) = 315\, [ppm] is the reference concentration in the year t0=1950t_0 = 1950.

In summary, we get the following parametrization of our outgoing longwave radiation source term

SOLW(T,x,t,CO2)=(A+B (T273.15)ΔA(CO2)), S_{OLW}(T,x,t,CO_2) = -(A + B\, (T - 273.15) - \Delta A(CO_2)),

which we reformulate into the shorthand notation

SOLW(T,x,t,CO2)=(A(CO2)+B T), S_{OLW}(T,x,t,CO_2) = - (A(CO_2) + B\,T),

where

A(CO2):=210.35.35 ln(CO2/315)) [W/m2], A(CO_2) := 210.3 - 5.35\,\ln(CO_2/315))\,\,[W/m^2],

and

B=2.15 [W/m2/K], B = 2.15\,\,[W/m^2/K],

where we further made the assumption that the unit of the temperature TT is [C][^\circ C] instead of Kelvin.


Created by Gregor Gassner and Andrés Rueda-Ramírez with contributions by Simone Chiocchetti, Daniel Bach, Sophia Horak, Philipp Baasch, Benjamin Bolm, Erik Faulhaber, and Luca Sommer. Last modified: April 02, 2026. Website built with Franklin.jl and the Julia programming language.