Measuring Uncertainties

The conclusions outlined above represent, in most cases, a relatively broad consensus among researchers. However, it is important to emphasize that all global and most regional assessments to date can best be characterized as "best-guess," usually supplemented with some simple sensitivity tests such as impacts with and without CO2 fertilization or adaptation. Yet a consensus among best guesses does not imply that we fully understand the risks associated with climate change, even at global scales (see Chapter 1). True uncertainty analysis, which attempts to quantify the total uncertainty inherited from all individual sources of uncertainty and estimate probabilities of different outcomes, has generally been absent. In some cases, failing to consider uncertainties can lead to a false sense of confidence about projections. In other cases, simple sensitivity tests can overstate uncertainties, such as comparing results from models run with and without CO2 fertilization (i.e. 30% or 0% fertilization) when the true range of uncertainty likely lies in between.

In our opinion, the remaining key need in impact assessments is to better characterize uncertainty and risks. This may include incorporating new processes (such as ozone or pest damage) but will likely center on better understanding the processes already treated in current models. Accomplishing this task is likely beyond the means of any single research group, given that no group has the number or diversity of models needed to evaluate the full suite of uncertainty sources. A proven strategy for assessing uncertainty is thus to compare model outputs from different groups, so-called model intercomparison projects (MIPs), such as is commonly done with climate models (see Chapter 3).

Periodic literature reviews and syntheses, such as those by the IPCC, provide useful insight into uncertainties but are not true MIPs in two important respects. First, studies often differ widely in the specific questions they address, and as a result typically evaluate different outcomes, time scales, and spatial scales of interest. For example, many studies have used equilibrium scenarios of doubled CO2, while others have used transient climate change projections. In the former, the climate system has come to equilibrium with atmospheric CO2 and therefore tends to be warmer than climate at the time of CO2 doubling in a transient simulation, because the climate system takes decades to respond to changes in CO2. Doubled CO2 experiments are therefore difficult to interpret as projections for any particular year. A more straightforward disparity is that many studies examine only production impacts or global commodity prices while others calculate changes in number of malnourished.

A second major challenge in the absence of MIPs is that most studies do not examine all relevant sources of uncertainty, and even a collection of studies will often all treat some model components in the same way. For example, most existing global assessments employ the same economic trade model (BLS), so that uncertainties associated with the structure of the trade model cannot be fruitfully explored.

Only when multiple groups follow the same model experiment design can a sizable number of simulations for the same variable of interest, and a systematic evaluation of the main potential sources of uncertainty, be ensured. One of the main obstacles to implementing MIPs is the substantial amount of foresight, coordination, and resources required to run multiple combinations of state-of-the art climate, crop, and trade models. But such MIPs will be crucial if we are to make progress in measuring uncertainties in impact or adaptation assessments.

An alternative, although by no means a substitute, to MIPs is to represent uncertainty in each component with simple statistical models that are computationally much more efficient and can readily represent uncertainties. This approach is exemplified by Tebaldi and Lobell (2008), who attempted to compute probability distributions of climate change impacts on global average yield changes for maize, wheat, and barley production for 2030 (see Chapter 3).

With all of the potential sources of uncertainty, one may wonder whether it is really necessary to examine each equation in each model or if a few key equations deserve most of the scrutiny. In fact, evaluating each equation is likely neither feasible nor necessary, but which should we focus on? Some insight into this question can be gained by evaluating projections with individual factors varied one at a time over a plausible range of values. This type of sensitivity analysis was used by Lobell and Burke (2008) to evaluate sources of uncertainty for projections of yield losses by 2030 in developing world regions.

Uncertainties from four factors were considered in the study: projected temperature change from climate models, projected precipitation change, sensitivity of crops to warming (estimated using time series analysis), and sensitivity of crops to rainfall. In most regions uncertainties related to rainfall were surprisingly small relative to temperature - surprising because year-to-year rainfall variations can be so important to crop production. However, temperature trends are much larger relative to historical variability than rainfall, with mean temperature trends typically twice as big as historical standard deviation while rainfall trends were much smaller than historical variability. In particular, that study identified crop sensitivity to temperature as a key unknown, i.e. there was often a big difference between impacts using a low vs high estimate of temperature sensitivity. Thus, efforts to quantify and reduce impact uncertainties would be well served by a focus on crop temperature responses. Uncertainties in future rainfall trends were less important overall, but emerged as the critical factor in a few key crops, such as rice and millets in South Asia.

Renewable Energy 101

Renewable Energy 101

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable. The usage of renewable energy sources is very important when considering the sustainability of the existing energy usage of the world. While there is currently an abundance of non-renewable energy sources, such as nuclear fuels, these energy sources are depleting. In addition to being a non-renewable supply, the non-renewable energy sources release emissions into the air, which has an adverse effect on the environment.

Get My Free Ebook

Post a comment