Global cooling from plant transpiration

How that could work

In recent decades, natural forests have continued to be decimated. Current climate models do not consider the post-deforestation reduction in transpiration as a significant climate change process. To get people think more seriously about what we are all losing, here we will attempt to advance the argument about a significant warming from reduced transpiration. To this end, one would need to do (at least) the following:

1. Provide a qualitative explanation of how reduced transpiration could cause global warming

2. Provide an order-of-magnitude estimate showing that the effect can be significant

3. Indicate how it can be assessed from observations

4. Show that observations support the prediction

5. Demonstrate that the global climate models (GCMs) show a mismatch with those observations (this is because we know that GCMs do not show any significant global effect from reduction in transpiration)

This discussion started as an email to EcoRestoration Alliance. It would be great to involve all our knowledgeable colleagues to a meaningful discussion and criticism of these propositions. It really matters.

1. Provide a qualitative explanation of how reduced transpiration could cause global warming

All solar energy absorbed by the Earth's surface is ultimately returned to space as the long-wave radiation. The energy captured by evaporation and transpiration at the surface is ultimately released to space from the mean height of condensation (in the tropical atmosphere, it is about 5-6 km). Incidentally (or not), this is also the mean effective radiating height of the Earth's atmosphere as a whole. From this height the long-wave radiation goes to space relatively unimpeded.

If we reduce transpiration, the extra solar energy T formerly spent on transpiration can now be converted to the long-wave radiation at the surface.

Now, importantly, instead of being radiated to space from the mean height of condensation (about 5-6 km), this energy is radiated into the atmosphere from near the surface. However, since the atmosphere has greenhouse gases, it only transmits to space a certain share k<1 of surface radiation.

So, the reduction in transpiration and the corresponding reduction in precipitation (precipitation and evaporation are always matched on a time scale of a few days) can lead to a reduction of the outgoing long-wave radiation by (1-k)T.

Since the incoming solar energy remains the same, while the outgoing radiation has decreased, the planetary surface begins to warm and emit more and more to space. This process stops when the extra emission to space matches the original imbalance of (1-k)T.

To estimate k, the share of surface radiation emitted to space, we consider that the Earth’s surface absorbs 168 W/m2 of solar radiation but emits up into the atmosphere 390 W/m2 (this is Stefan-Boltzmann radiation for the global mean surface temperature of 288 K), which corresponds to k ~ 0.4.

2. Provide an order-of-magnitude estimate showing that the effect can be significant

During the industrial era, forests have been destroyed on approximately 13% of land area. Assuming that deforestation reduces latent heat flux by 30% from its global mean value of 80 W/m2, and taking into account that land constitutes 0.29 of total planetary surface, we obtain a reduction of T=0.13*0.3*80*0.29 W/m2 = 0.9 W/m2.

With k ~ 0.4, we obtain a radiative forcing of (1-k)T ~ 0.5 W/m2.

In comparison, CO2 forcing has grown to about 2 W/m2 from the beginning of the industrial era till 2020. This means, conservatively, that evapotranspiration reduction could have so far accounted for about one fifth of global warming.

3. Indicate how it can be assessed from observations

In brief, we should check temperature trends at different heights. This is because radiation is determined by temperature, so if the radiative fluxes change, the corresponding temperatures must change too.

In our thought experiment, the effective radiative height (~ 5 km) does not change, because the amount of absorbers does not change (CO2 is fixed). By reducing transpiration, we effectively reduce the radiative flux at the effective radiative height and increased it at the surface. When the new steady state is reached, the surface will radiate more, while the upper radiative layer will radiate as before matching the solar absorption. So the difference between the surface upward flux of thermal radiation and the flux of outgoing thermal radiation has increased.

This means that the temperature trend should be zero at the effective radiative height ~ 5 km and increasingly positive at lower altitudes.

In contrast, if the surface warming is due to increased CO2 with no evaporation reduction, the effective radiative height increases (by about 120 m for CO2 doubling). This is because there is more CO2 in the atmosphere, and the upper radiating amount is now displaced higher. Then there is uniform warming at all heights including the initial effective radiative height of 5 km.

Two simple schemes illustrate how it works (I explain this in greater detail here).

In the graph above, the black horizontal line indicates the effective radiative height that increased from its original value of 5 km (the gray line) due to a(n unrealistically high) CO2 increase. Note that at 5 km the warming is the same as it is at the surface (the rosy and red lines are parallel).

In the graph below, the atmosphere radiates to space from the same height as before (5 km), therefore its temperature at this height must be the same (for the outgoing radiation to match the incoming solar radiation). 

If we have both processes at work simultaneously, CO2 increase and transpiration reduction, then the temperature will grow at around 5 km but more slowly than in the lower atmosphere. The more slowly the temperature grows at 5 km compared to how it grows at the surface, the greater proportion of warming is due to the evaporation reduction.

4. Show that observations support our prediction

5. Demonstrate that the global climate models show a mismatch with those observations  (this is because we know that GCMs do not show any significant global effect from reduction in evapotranspiration)

Now let us turn to the available data. In the graph below we can see the difference between surface temperature trends dTs/dt and the temperature trends at different atmospheric levels according to models (amip, cmip) and observations (ERA5, AIRS), d(Ts - T)/dt.

We can see that in all the models the temperature trends at all heights are the same or larger than at the surface, so the model curves tend to the negative region of the horizontal axis. (This reflects the well-known decrease of the moist adiabatic lapse rate with increasing temperature).

In contrast, in observations, especially (expectedly) on land, we can see something different. In particular, at the effective radiative height (5 km approximately corresponds to 500 hPa) we can see that air temperature increases much more slowly than the surface temperature.

For example, for ERA5 the surface temperature increases at a rate of 0.28 K/decade, while at 500 hPa it increases at a rate of only 0.12 K/decade, such that d(Ts-T)/dt = 0.16 K/decade as shown in the graph for land.

This ratio is consistent with our idea that a considerable part of the warming is due to changes in the lapse rate (and ultimately with transpiration/precipitation reduction). 

It should be noted that in the climate literature the lapse rate trends observations have always been characterized as "controversial" -- now for more than twenty years.

Again, it would be great if these ideas were carefully scrutinized, criticized and discussed and if they stand a careful scrutiny one could think of developing simple and captivating messages on this basis showing how forests really matter.