by Douglas Sheil
Despite major investments in incorporating land cover in climate simulation models, much remains uncertain, especially concerning the influence of land-cover change on cloud cover and rain.
It’s been over 16 months since our last guest blog on condensation-driven winds. That blog generated plenty of feedback so I hope that you may be interested in an update. We have published a number of papers further refining our ideas (see, e.g., here and here) and have working on several more (see, e.g., preprints here and here).
But here my aim is not to reopen debate on the technical details of the condensation theories, but rather to highlight advances and implications of more general interest. I focus on two papers (and highlight a few more at the end): one summarizes some recent advances in studying the links between rain and vegetation (here, free online until the end of June I believe) and another that examines how condensation and atmospheric pressure changes are related to forest cover (here, behind a paywall but also available at ArXiv). As the first paper is written for a general audience most space is devoted to summarizing the key message from the second paper.
Paper 1: “How plants water our planet”
The first paper is short and targeted at a biological readership. It argues that biology plays a much greater role in determining rainfall, and climate generally, than is widely recognized. It attempts to summarize recent advances that illustrate this argument and its importance.
The article notes that “despite major investments in incorporating land cover in climate simulation models, much remains uncertain, especially concerning the influence of land-cover change on cloud cover and rain”. This is not controversial. As one recent commentary on climate models notes explicitly rainfall over land remains hard to simulate because it is largely determined by ‘unresolved processes’. This represents the ‘main limitation in current representations of the climate system’ and ‘a major roadblock to progress in climate science’ (Science 340, 1053-1054).
I focus on rain, but the implications of the terrestrial water cycle are bigger. If you are interested in temperature it is worth noting that the vaporization of water consumes nearly half the solar energy reaching the Earth’s land surface and contributes to local cooling (and the redistribution of heat). Atmospheric moisture also impacts the climate in many ways: water vapour is a powerful ‘greenhouse gas’ and cloud cover influences planetary albedo (a measure of the solar radiation reflected into space).
One recent study (Nature 496, 347-350) estimates that transpiration (water vapour derived from plants) produces 80–90% of the atmospheric moisture derived from continents. This figure substantially surpasses previous estimates (20–65%). If the share of atmospheric moisture derived from vegetation is so much larger than previously recognised, then changes in vegetation may also have greater impacts. Until the processes underlying vegetation control of the water cycle are resolved, the potential impact of land-cover change on the regional and global temperature regimes cannot be estimated with confidence.
The review highlights some further advances describing how trees and forests may influence the global water cycle and associated ‘unresolved processes’. The biotic pump idea (see below) is only one of these exciting advances, though admittedly it appears among the most profound and controversial. Among other things I speculate that both forest loss and increasing atmospheric carbon dioxide may help explain a reduction in tropical winds, and changes in regional weather patterns.
I also suggest that changes in Walker Circulation may be influenced by the rapid forest loss occurring in South East Asia.
It is not controversial to suggest that we need a much better understanding of how land-cover influences climate. There are major gaps in our understanding. As Anastassia and Victor recently commented to me (when I shared a draft of this blog), much of what is currently considered “natural climatic variability” might in fact reflect the effects of human land-cover change if only we adequately understood these influences better. (Historically many regions have been subjected to major forest loss, and/or regrowth). This is one reason why it is crucial to take forests, and land cover more generally, into account in climatic reconstructions.
Please read the paper and judge for yourselves. It provides context for our second paper.
Paper 2: “Why we must reassess the role of forests for climate and for water security”
A commentary on Makarieva A.M et al. 2014 Journal of Hydrometeorology, 15, 411-426.
This is derived from a text that Anastassia and the team developed together (see here). I share it here, in slightly edited form, with their permission.
The effects of forests on weather are often viewed solely in terms of moisture recycling. That is, evaporation from forest returns moisture to the atmosphere where it can increase local rain. In our new paper, using a recent study as an example (Nature 489, 282–285), we discuss limitations in this perspective. Among other things, we argue that the existing global circulation models cannot be used for evaluation of the climatic effects of deforestation because of their inability to reproduce various phenomena that we have ascribed to a new mechanism we have termed the “biotic pump” – by this mechanism forests generate large-scale pressure gradients that cause winds to flow and bring moisture from oceans to land.
Below we explain some highlights from the second part of our study that examines the pressure relationships that our theories predict. Here we have tried to summarise and illustrate the key ideas for a general readership.
Moist air arrives to land in the lower atmosphere, rises and becomes depleted of moisture (as clouds form and rain falls). This dry air returns to the ocean in the upper atmosphere while the net imported moisture returns to the ocean via rivers (and other minor flows). Air circulation models describe the aerial component of the water cycle: i.e. how much moisture is carried by winds. Measures of river runoff provide an independent check on any such models’ validity as this runoff must match what the winds bring in. (Unfortunately the circulation models escape any such check in the oceans as there is no runoff to measure).
Current air circulation models do not pass this check: inputs don’t match outputs. For example, the Amazon models can only account for half the measured river runoff. Similar considerable discrepancies are common for all regional models and no fix to this problem has yet been identified (e.g., J. Hydrometeor 12, 556–578).
In our recent paper we provide new evidence for the biotic pump. We analyse the relationship between wind direction and surface air pressure in forested and deforested regions of the Amazon basin. This highlights how the intense evaporation from forest creates low pressure. Let’s take a look – schematically – how this condensation-evaporation cycle works.
Immediately after rain the local atmosphere is relatively dry (water vapour has condensed and precipitated). Winds are negligible. The atmosphere slowly regains water vapour via evaporation.
Owing to the high cumulative surface of leaves, the forest enriches the atmosphere with water vapour more rapidly than does the ocean. Total air pressure in the area slowly grows reflecting the accumulating water vapour. (Our analyses show that rainy days in the Amazon forests are on average characterized by a slightly higher pressure than the rainless days. In the deforested region the opposite is true – we discuss this is greater detail in the paper (see Section 4b,c), but the pattern is consistent with the biotic pump.)
Rainfall probability as a function of the total amount of water vapour in the atmosphere (CWV). As used in Fig. 5 of Makarieva et al. (2014) based on the data of Holloway and Neelin (2010).
Small differences in evaporation rates translate into large differences in the probability of rainfall due to the sharply rising relationship between water vapour content and the likelihood of rain. Thus rain is much more likely to start again over the forest.
Once sufficient water vapour has accumulated over the forest, condensation begins. Local air pressure starts to decline. In contrast to the gradual process of water vapour evaporation, condensation can occur quickly.
The resulting pressure differences now draw wet air from the ocean to the forest, this air rises and cools. Now moisture generated locally via forest evaporation precipitates on land together with additional moisture brought from the ocean. This additional moisture is what ensures land remains wet while rivers keep running back to the ocean.
Thus, rather than merely recycling moisture, forests actually drive the water cycle on land. Recognition of this role will lead to re-evaluation of the importance of natural forests and the need for forest conservation to prevent water scarcity. We urge a major reassessment of the role of forests in atmospheric dynamics.
One interesting study is provided by German Poveda, Liliana Jaramillo, and Luisa Vallejo, who have used the biotic pump ideas to interpret the behaviour of the so-called “Caribbean Low Level Jet” which curves towards South America against the prevailing trade winds. The observations are consistent with forest generating low pressure to draw in winds. (Anastassia and Victor offer further comment here ).
Another interesting paper is by Mark Andrich and Jörg Imberger who seek to distinguish rainfall changes in Western Australia caused by land-cover and more general shifts in the Hadley circulation. Their conclusion is that land-cover (i.e. tree loss) is the dominant factor to explain precipitation changes. (See also here).
For more information see
Sheil D. (2014) How plants water our planet: advances and imperatives. Trends in Plant Science 19, 209-211 doi: 10.1016/j.tplants.2014.01.002
Makarieva A.M., Gorshkov V.G., Sheil D., Nobre A.D., Bunyard P., Li B.-L. (2014) Why does air passage over forest yield more rain? Examining the coupling between rainfall, pressure, and atmospheric moisture content. Journal of Hydrometeorology, 15, 411-426, doi:10.1175/JHM-D-12-0190.1. Abstract.Supplemental information.
JC note: This is a guest post submitted to me by Douglas Sheil. As with all guest posts, keep your comments relevant and civil; moderation will be heavier than usual.