With a growing number of people, we will have to produce more and more food in the future. This is only possible if we have enough water for agriculture. We need to know whether there will be more or less water in the future – that is, whether soil moisture will change. To estimate future changes, we also have to understand past changes and what the reasons for these changes were.
For our recent article (Mueller and Zhang, 2016) published in the journal Climatic Change we analyzed changes in soil moisture levels over 1951-2005. We focused on land areas in the northern hemisphere. In a nutshell, the areas have become drier. This drying cannot be explained by natural variability of the climate system, but is likely caused by increasing greenhouse gas concentrations.
Drying trends found across different data sources
A general drying of the land surfaces was found in several independent soil moisture estimates:
- In Satellite measurements (ESA-CCI): These data are not shown here because they cover a shorter time-period. However, they do show a drying over 1979-2008 as you can see in our article.
- A blend from observations and a model, so-called land-surface model estimates.
- Model simulations (CMIP5) under realistic assumptions.
All three estimates suggested that land areas got drier. The green bar in the figure below is the trend from the blend from observations and model, and the red bar from the model simulations under realistic assumptions.
Opposite trends without greenhouse gas emissions
We also compared the blend from observations and model with models that estimate soil moisture under the assumption that greenhouse gas concentrations did not change over 1951-2005 – in other words from models that have only natural forcing. These models show a wetting trend rather than the observed drying as seen in the blue bar below.
Detecting our influence on soil moisture changes
The fact that models that do not include changes in CO2 and other man-made climate forcing do not agree with observations is an indication that man-made climate forcing may be important for soil moisture changes. We tested this hypothesis in a linear regression framework (regularized optimal fingerprinting). The principle of the method is explained in an earlier article by thedatajournalist using the example of temperature rather than soil moisture. We calculated space-time changes of soil moisture and determined the agreement between the soil moisture estimates from observations and models. The result is shown in the figure below. The red bar is the result for the comparison of observations with models with forcing from natural and man-made sources, including CO2. The entire bar is above zero, which means man-made climate forcing had an influence on changes in soil moisture. It also means that changes in soil moisture cannot be explained by internal variability – such as El Nino or other oscillations – alone. The blue bar represents the result for natural influences only. It is not larger than zero, and we say that the ‘responses to natural forcing alone cannot be detected in the observations’. The figure thus shows that while natural climate forcing alone cannot explain changes in soil moisture, anthropogenic (man-made) influences together with natural influences can explain the drying trends. We conclude that our past emissions of greenhouse gases and aerosols may likely be the reason why land areas have become drier in the past. This is a strong simplification of our article, and more thorough analyses with sensitivity tests can be found in the original article. In the article, we also found similar results for the size of areas under drought conditions: They got larger over the past 55 years, and this can likely be explained by our impact on the climate system.
As an important indicator of water availability, changes in soil moisture can have significant implications for agriculture and food security. We demonstrated a human contribution to the increase in dryness in the northern hemisphere. The drying is just one of the many changes in the climate system that result from our emissions of greenhouse gases.
What the future holds
What the future will bring us is difficult to tell. But even though we do not have measurements for the future, we have future estimates from model simulations, and the comparison of past model simulations with observations gives us confidence in these models. Furthermore, with increasing temperatures, drier soils might be expected due to physical mechanisms (see Section 1). Future simulations from climate models indeed show increasing drought conditions throughout the 21st Century, as seen in the video below.
It seems that a dry future lies ahead of us.
Video: Palmer drought severity index, CMIP3, NCAR Vislab.
Mueller and Zhang: Causes of drying trends in northern hemispheric land areas in reconstructed soil moisture data, Climatic Change, 2016. http://link.springer.com/article/10.1007%2Fs10584-015-1499-7