How Do Scientists Know Human Activities Impact Climate? A brief look into the assessment process

Written by Levi Golston

On the subject of climate change, one of the most widely cited numbers is that humans have increased the net radiation balance of the Earth’s lower atmosphere by approximately 2.3 W m-2 (Watts per square meter) since pre-industrial times, as determined by the Intergovernmental Panel on Climate Change (IPCC) in their most recent Fifth Assessment Report (AR5). This change is termed radiative forcing and represents a basic physical driver of higher average surface temperatures resulting from human activities. In short, it elegantly captures the intensity of climate change in a single number – the higher the radiative forcing, the larger the human influence on climate and the higher the rate of increasing surface temperatures. Radiative forcing is also significant because it forms the basis of equivalence metrics used in international environmental treaties, defines the endpoint of the future scenarios commonly used for climate change simulations, and is physically simple enough that it should be possible to calculate without relying on global climate models.

Given its widespread use, it is important to understand where estimates of radiative forcing come from. Answering this question is not straightforward because AR5 is a lengthy report published in three separate volumes. Chapter 8 of Volume 1, more than any other, quantitatively describes why climate change is occurring due to natural and anthropogenic causes and is, therefore, the primary source for how radiative forcing is assessed by the IPCC. One of the key figures is reproduced below, illustrating that the basic drivers of climate change are human-driven changes to aerosols (particles suspended in the air) and greenhouse gases, along with their relative strengths and uncertainties:

Fig. 1: Assessments of aerosol, greenhouse gas, and total anthropogenic forcing evaluated between 1750 and 2011. Lines at the top show the 5-95% confidence range, with a slight change in definition from AR4 to AR5 [Source: Figure 8.16 in IPCC AR5].
This post seeks to answer two questions: how is the 2.3 W m-2 best estimate determined in AR5? And further, why is total anthropogenic forcing not known more precisely than shown in Figure 1 given the numerous observations currently available?

1. Variations on the meaning of radiative forcing

Fundamental laws of physics say that if the Earth is in equilibrium, then average temperature of the Earth is such that there is a balance between the energy that the Earth receives and the energy that it radiates. When this balance is disturbed, the climate will respond due to the additional energy in the system and will continue to change until the forcing has fully propagated through the climate system at which point a new equilibrium (average temperature) is reached. This response is controlled by processes with a range of timescales (e.g. the surface ocean over several years and glaciers over many hundreds of years), so radiative forcing depends on when exactly it is calculated. This leads to several subtly differing definitions. While the IPCC distinguishes between radiative forcing and effective radiative forcing, I do not attempt to distinguish between the two definitions here and refer to both as radiative forcing.

Figure 2 shows the general framework for assessing human drive change used by the IPCC, which is divided into four major components. Firstly, the direct impact of human activities through the release (emission) of greenhouse gases and particulates into the atmosphere is estimated, along with changes to the land surface through construction and agriculture. These changes cause the accumulation of long-lived gases in the atmosphere including carbon dioxide, the indirect formation of gases through chemical reactions, and an increase in number of aerosols in the atmosphere (abundance). Each of these agents influences the radiation balance of the Earth (forcing) and over time causes warming near the surface (climate response).

Fig. 2: Linear [uncoupled] framework for modeling climate change shown with solid arrows. Dashed arrows indicate climate feedback mechanisms driven by future changes in temperature. Credit: Levi Golston
2. Individual drivers of change

The two major agents (aerosols and greenhouse gases) are further sub-divided by the IPCC as shown below. Each of the components are assessed independently, then summed together using various statistical techniques to produce the best estimate and range shown in Figure 1.

Fig. 3: Estimates of radiative forcing (dotted lines) and effective radiative forcing (solid lines) for each anthropogenic and natural agent considered in AR5. [Source: Figure 8.15 in IPCC AR5].
Since the report itself is an assessment, each of the estimates in Figure 3 were derived directly from the peer-reviewed literature and are not the result of new model runs or observations. I have recently identified the specific sources incorporated in this figure elsewhere if one wants to know exactly how any of the individual bars were calculated. More generally, it can be seen that the level of confidence varies for each agent, with the most uncertainty for aerosol-radiation and aerosol-cloud interactions. Positive warming is driven most strongly by carbon dioxide followed by other greenhouse gases and ozone. It can also be seen that changes in solar intensity are accounted for by the IPCC, but are believed to be small compared to changes from the human-driven processes.

3. Can net radiative forcing be more directly calculated?

Besides adding together individual processes, is it also possible to independently assess the total forcing itself, at least over recent decades where satellite and widespread ground-based observations are available?  In principle, changes in the Earth’s energy balance, primarily seen as reduced thermal radiation escaping to space and as heat uptake by the oceans, should relate back to the net forcing causing these changes in a way that provides an alternate means of calculating human’s influence on the climate. To use this approach, one would need a good idea of how sensitive the Earth’s response will be in response to a given level of forcing. However, this sensitivity is equally or more uncertain than the forcing itself, making it difficult to improve on the process-by-process result. The ability to account for the Earth’s overall energy balance and then quantify radiative imbalances over time also remains a challenge. Longer data records and improved knowledge of climate sensitivity may eventually advance the ability to directly determine total radiative forcing going forward.

Fig. 4: Simulation of CO2 concentrations over North America on Feb 12th, 2006 by an ultra-high-resolution computer model developed by NASA. Photo Credit: NASA
4. Summary

The most widely cited number is based on an abundance-based perspective with step changes for each forcing agent from 1750 to 2011, resulting in an estimated total forcing of 2.3 W m-2. This number does not come from an average of global climate models, as might be imagined, but instead is the sum of eight independent components (seven human-driven, one natural), each derived and assessed from selected recent sources in the peer-reviewed literature.

Radiative forcing is complex and requires models to translate how abundances of greenhouse gases and aerosols actually affect global climate. For gases like carbon dioxide, documented records are available going back to pre-industrial times and earlier, but in other cases additional modelling is needed to help determine the natural state of the land surface and atmosphere. The total human-driven radiative forcing (Figure 1) is still surprisingly poorly constrained in AR5 (1.1 to 3.3 W  m-2 with 90% confidence), which is a reminder that while we are certain human activities are causing more energy to be retained by the atmosphere, continued work is needed on the physical science of climate change to determine by exactly how much.


Levi Golston is a PhD candidate in Princeton’s Environmental Engineering and Water Resources program. His research develops laser-based sensors coupled with new atmospheric measurement techniques for measuring localized sources, particularly for methane and reactive nitrogen from livestock, along with methane and ethane from natural gas systems. He blogs at

Leave a Reply

Your email address will not be published. Required fields are marked *