Introduction

Main publication
(abstract)

Overview of the results

Comments and responses
[you are here]
List [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

Savenije H. (2006) Interactive comment on "Biotic pump of atmospheric moisture as driver of the hydrological cycle on land" by A. M. Makarieva and V. G. Gorshkov. Hydrology and Earth System Sciences Discussions, 3, S993-S996. www.cosis.net/copernicus/EGU/hessd/3/S993/hessd-3-S993.pdf
The comment discusses the main biotic pump publication.
Informal title of the comment:
Comment on Biotic Pump

This is an interesting paper, that I hope will trigger ample discussion. I cannot fully judge the meteorological part of the paper, but I want to comment on the moisture recycling which is an important factor in sustaining continental rainfall.

Makarieva A.M., Gorshkov V.G. (2006) Interactive comment on "Biotic pump of atmospheric moisture as driver of the hydrological cycle on land" by A. M. Makarieva and V. G. Gorshkov. Hydrology and Earth System Sciences Discussions, 3, S1005-S1011. www.cosis.net/copernicus/EGU/hessd/3/S1005/hessd-3-S1005.pdf

In response to the comments by H. H. G. Savenije: We are very grateful for the critical reading of our paper and for the constructive comments that raise important issues. Below we adress each of the numbered comments in their original order.

1. The authors write that forest evaporation is (mostly) transpiration, but this is not true. On page 2629 they write "When the soil moisture content is sufficiently high, transpiration, which makes up a major part of total evaporation from the surface, E, is dictated by solar energy." However, evaporation from interception is as important, or even more important than transpiration in this case. In tropical forests evaporation from interception can amount to more than 50% of the total evaporation (Savenije, 2004), and also in temperate forests (Gerrits et al. 2006). Interception has less resistance to evaporation than transpiration, and even less than open water evaporation. Only when you consider the important contribution of interception (both canopy and forest floor interception) to the total forest evaporation, can you make it plausible that forests evaporate more than open water or the sea. So in fact mentioning that interception accounts for about 50% of the evaporation only strengthens your case. You could say, "When the soil moisture content is sufficiently high, transpiration is dictated by solar energy. Interception, both from the canopy, understorey and the forest floor, which can be more than 50% of the total evaporation (see Savenije, 2004), is also dictated by solar energy, with a much lower resistance to evaporation than transpiration or even open water evaporation. So if there is sufficient soil moisture, then evaporation from dense forest is merely constrained by the energy balance, driven by solar radiation, which can be more efficient than sea water evaporation"

1) We agree that interception, this important term in total forest evaporation, should be discussed on p. 2629, as suggested in the comment. The reason why originally we focused on transpiration is because transpiration, unlike interception, is under immediate, short-term biotic control. For example, as we mention in the paper (p. 2650-2651), closure of the stomata in the afternoon results in a drop of the total evaporation flux. As a consequence, the evaporative force diminishes and the upwelling cloud-supporting air flux diminishes as well, leading to formation of precipitation. The daily opening/closure of the stomata and the associated growth/decrease of the evaporative force can be responsible for the diurnal cycle of precipitation in the tropics (Nesbitt et al. 2003), which, remarkably, is observed on land but not in the ocean, where the stomata effects are obviously absent.

Interception, on the other hand, is also dictated by such biotic properties as are, e.g., leaf area index or leaf size. However, these properties, originated in the course of biological evolution, can be only regulated in the long-term, e.g. on a seasonal scale. For a given forest on a short-term scale, the interception term is therefore dictated by physical, rather then biological, regularities: the intercepted water evaporates with no further biotic control. This leads to an interesting idea. Evergreen coniferous forests in the boreal zone retain high leaf area index in winter. The associated high interception of winter precipitation enhances total evaporation thus making biotic pump working even in winter forests where no photosynthesis or other biological processes occur. In other words, being evergreen helps the trees to enhance winter precipitation. This might explain the wide spread of evergreen trees (pines, firs, spruces) in the pristine boreal forests. Secondary boreal forests are composed of deciduous species that lack this feature.

It is remarkable that forest transpiration and interception are close in magnitude and apparently are both significant for the biotic pump functioning. Recently Cuartas et al. reviewed the available interception estimates for Amazonian forests to find that interception ranges from 9 to 26% of total rainfall. If one assumes that runoff, on average, constitutes approximately 50% of Amazonian precipitation (Marengo, 2005), this means that interception constitutes from about 20 to more than 50% of total evaporation, as is also the case in other ecosystems (Savenije, 2004). Since we do not discuss the resistance to evaporation of the various total evaporation constituents, we would like to modify the text change on p. 2629 (line 10), that was suggested in the comment, as follows:

"When the soil moisture content is sufficiently high, transpiration is dictated by solar energy. Interception, both from the canopy, understorey and the forest floor, which can be more than 50% of the total evaporation (see Savenije, 2004), is also dictated by solar energy. So if there is sufficient soil moisture, then total evaporation from dense forest is constrained by solar radiation. Therefore, when x is counted along the parallel, on well-moistened continental areas we have E(x) = E(0), i.e. evaporation does not depend on the distance from the ocean."

2. I am not sure that rainfall is proportional to the flux. Rather to the moisture content over a threshold (see Savenije, 1996b). But this may not be very important since it results into the same equation (3). In the theory on moisture recycling described by Savenije (1995), which the authors refer to, the decrease of the precipitation along the trajectory of the moisture transport is caused by the runoff and the recharge to groundwater (see Savenije, 1996a). Only if there is 100% moisture recycling can there be constant rainfall along a trajectory. Here lies the main difference between the work of the biotic pump and the theory on moisture recycling. The authors in fact imply that the biotic pump is able to compensate for the drainage of moisture by runoff.

2) To derive the exponential equation (3), p. 2626, for precipitation decline with distance from the ocean one has to make certain assumptions. In Savenije (1995) the exponential equation for precipitation decline ultimately stems from the assumed proportionality between precipitation, on the one hand, and the product of precipiable atmospheric water and air stream velocity (not precipiable moisture content alone), on the other. This assumption is equivalent to our assumption of proportionality between precipitation and flux F in our paper, because flux F is actually equal to moisture content in the atmospheric column multiplied by air stream velocity (which greatly changes with distance).

However, the theoretical question why precipitation over deforested territories declines exponentially with distance, is not of primary concern in our paper. It is enough for our considerations that empirical data strongly confirm this relationship in sharp contrast with natural forests.

The comment says: "In the theory on moisture recycling described by Savenije (1995), which the authors refer to, the decrease of the precipitation along the trajectory of the moisture transport is caused by the runoff and the recharge to groundwater (see Savenije, 1996a). Only if there is 100% moisture recycling can there be constant rainfall along a trajectory. Here lies the main difference between the work of the biotic pump and the theory on moisture recycling. The authors in fact imply that the biotic pump is able to compensate for the drainage of moisture by runoff." We would like to add that in the theoretical considerations of moisture recycling precipitation P is ultimately related to the horizontal flux of moisture, which is considered as an abiotic geophysical constraint, an independent parameter; evaporation E is known to be affected by vegetation; and runoff R is the residual of P and E. This consideration, where both P and R can be arbitrarily high or low, corresponding to either droughts and floods, should hold well for abiotic environments like deserts, agricultural lands etc.

In our biotic pump consideration we assign a different meaning to the budget P = E+R. In the biotically controlled forest environment runoff R is determined by the biotically maintained high soil moisture, total evaporation E is dictated by solar energy, and precipitation P is biotically regulated (via the biotically regulated flux F) to balance the equation. In this consideration it is clear that, as is also confirmed by observations, in natural forests neither floods nor disastrous diminishment of runoff can normally occur throughout the year.

3. The reasoning on the top of page 2629 is not correct (although it does not necessarily affect your conclusion from the reasoning). The authors state: "High stationary (dW/dt=0) soil moisture content is inseparably associated with a significant runoff, i.e. loss of water by the ecosystem." This is not true. What you should say is: "A high amount of soil moisture (average over the year) implies significant runoff, i.e. loss of water by the ecosystem (because R is a function of W)". One should realise that maintaining a high moisture content does not mean that dW/dt=0 (stationarity). At the short temporal scales dW/dt is never zero, it varies with the difference in phase of P, E and R, but at the average annual scale, you can indeed say that stationarity implies that dW/dt=0. However, this does not mean that the annual average W is high and constant along the x-axis. The condition of stationarity can apply to both a low and a high average moisture level. What the authors mean with stationarity is that in the long run (over several years) P=E+R (the long term average of dW/dt=0). Although this is true at any point along a trajectory, it does not mean that dW/dx=0. The situation can be stationary and still have a longitudinal gradient. What you can say is that if dW/dx=0 then dR/dx should be zero as well, because R=f(W). Subsequently, the fact that R=P-E implies that d(P-E)/dx=0. This in fact implies that the difference between P and E is constant. One solution is that both P and E are constant along the trajectory. And this is what you conclude. But a steady state solution also occurs if the difference between P and E is constant, while both are decreasing. However, you indeed observe that in some cases (Amazon and Congo) dP/dx=0, and that then indeed implies (in your reasoning where dR/dx=0) that also dE/dx=0. In the case of the Yenisey, however, you have dP/dx>0, which means that also dE/dx must be larger than zero, increasing along a trajectory. So it is not as simple as the authors state in the first paragraph of page 2629.

3) By the term dW/dt = 0 on the top of page 2629 we indeed meant its long-term average, in agreement with the formulation suggested in the comment: "A high amount of soil moisture (averaged over the year) implies significant runoff, i.e. loss of water by the ecosystem (because R is a function of W)". We also agree with all further derivations in the comment and in fact it is precisely what is said in the paper in the two upper paragraphs on p. 2629, which should be jointly considered. Indeed, what we say is that if dW/dx = 0, then dR/dx = 0 ("In areas where neither the surface slope, nor soil moisture content depend on distance x from the ocean, W(x) = W(0), loss of ecosystem water to runoff is spatially uniform as well, R(x) = R(0).")

Since P = E +R, dR/dx = 0 means, as indicated in the comment, that d(P - E)/dx = 0. In the second paragraph on p. 2629 it is essentially said that E in forests is dictated by solar energy, so when x is counted along the parallel, dE/dx = 0, while when x is counted along the meridian then dE/dx > 0. From this and d(P - E)/dx = 0 it is concluded that along the parallel dP/dx should be zero, but it must be positive when x is counted towards the equator (in forested areas).

The concluding lines of the first paragraph on p. 2629, "It follows that in the stationary case precipitation P, which maintains stationary soil moisture content and compensates the runoff, cannot decrease with distance from the ocean either. The conditions W(x) = W(0) and R(x) = R(0) are incompatible with an exponential decline of P(x), Eq. (3).", which disrupt the logic of the above consideration with the premature conclusion about constant P made from dR/dx = 0 alone, should be perhaps deleted or moved to the end of the second paragraph.

We would finally formulate the first two paragraphs on p. 2629 as follows (with account of comment 1):

"Change of soil moisture content with time, dW/dt, is linked to precipitation P, evaporation E and runoff R via the law of matter conservation, dW/dt = P - E - R. A high amount of soil moisture (averaged over the year) implies significant runoff, i.e. loss of water by the ecosystem (because R is a function of W). In areas where neither the surface slope, nor soil moisture content depend on distance x from the ocean, W(x) = W(0), loss of ecosystem water to runoff is spatially uniform as well, R(x) = R(0).

When the soil moisture content is sufficiently high, transpiration is dictated by solar energy. Interception, both from the canopy, understorey and the forest floor, which can be more than 50% of the total evaporation (see Savenije, 2004), is also dictated by solar energy. So if there is sufficient soil moisture, then total evaporation from natural forest is constrained by solar radiation. Therefore, when x is counted along the parallel, where the amount of the incoming solar radiation does not change, on well- moistened continental areas we have E(x) = E(0), i.e. evaporation does not depend on the distance from the ocean. Coupled with constant runoff, R(x) = R(0), this means that precipitation P is similarly independent of the distance from the ocean, P(x) = E(0) + R(0) = P(0). When the considered area is oriented, and x counted, along the meridian, evaporation increases towards the equator following the increasing flux of solar energy. In such areas, provided soil moisture content and runoff are distance-independent, precipitation must also grow towards the equator irrespective of the distance from the ocean. The conditions W(x) = W(0) and R(x) = R(0) are incompatible with an exponential decline of P(x), Eq. (3)."

4. I do not understand completely how the biotic pump works. I am not a specialist in meteorology and atmospheric circulation, but with the Yenisey case, you certainly have a strong indication that something special is happening here. So from my point of view I give the authors' theory the benefit of the doubt.

5. When you write on the bottom of page 2630:" Thus, precipitation over forests increases up to the maximum value possible at a given constant runoff (i.e., coefficient k in Eq. (2), the precipitation/runoff ratio, is maximized for a given R.)." In Savenije 1996b, I called this coefficient k the "multiplier", or, in an earlier paper, "the return rate", the inverse of the runoff (Savenije 1996a). Indeed when the multiplier is large, then the recycling is large. With zero runoff, the multiplier is at its maximum (approaching infinity) and the recycling approaches 100%, resulting in constant rainfall along a trajectory (dP/dx=0) and P=E. I think that the term "multiplier" is also appropriate in your case.

6. The fact that you demonstrate that one can actually get constant or increasing P while there still is runoff, means that apparently there is an additional mechanism at work, besides recycling. If that is indeed your biotic pump, then you have made an important contribution to understanding the continental moisture balance.

4, 6) Since our paper was submitted to HESS, we have substantially extended our analyses of precipitation-distance patterns in the large forested regions of the world, including river basins of Mackenzie in North America, Lena and Ob in Siberia (Makarieva and Gorshkov, 2006; Makarieva, Gorshkov, Li, in preparation). In all these meridian-oriented regions we observed an increase of precipitation towards the inner part of the continent, similar to the situation with Yenisey. Moreover, the analysis has shown that this increase, as theoretically predicted, is well-matched in magnitude by the increase in solar radiation with decreasing latitude. As we state in the paper (p. 2628 (lines 6-11), p. 2651 (lines 9-13)), irrespective of the described physical details of functioning of the biotic pump, the analyzed observations unambiguously testify to its existence.

References:

Savenije, Hubert H.G., 2004. The importance of interception and why we should delete the term evapotranspiration from our vocabulary. Hydrological Processes, 18(8):1507- 1511.

Savenije, H.H.G., "Does moisture feedback affect rainfall significantly?", Physics and Chemistry of the Earth, Vol. 20, No. 5-6, pp.507-513, 1996b.

Savenije, H.H.G., "The Runoff Coefficient as the Key to Moisture Recycling", Journal of Hydrology, 176:219-225, Elsevier, Amsterdam, The Netherlands, 1996a.

Savenije, H.H.G., "New definitions for moisture recycling and the relation with land-use changes in the Sahel". Journal of Hydrology, 167:57-78, Elsevier, Amsterdam, The Netherlands, 1995.

Gerrits, A.M.J., Savenije, H.H.G. and Pfister, L. , 2006. Measuring forest floor interception in a beech forest in Luxembourg, Hydrology and Earth System Sciences Discussions, 3, 2323-2341.

References

Cuartas, L. A., Tomasella, J., Nobre, A. D., Hodnett, M. G., Waterloo, M. J. and Munera, J. C.: Interception water-partitioning dynamics for a pristine rainforest in Central Amazonia: Marked differences between normal and dry years, Agricultural and Forest Meteorology, in press.

Makarieva, A. M. and Gorshkov, V. G.: Rivers. Will they be ever flowing on Earth?, Ecology and Education, 1-2(2006), 7-11. (in Russian)

Makarieva, A. M., Gorshkov, V. G., Li, B.-L.: Precipitation on land versus distance from the ocean: Evidence for a forest pump of atmospheric moisture, in preparation.

Marengo, J. A.: Characteristics and spatio-temporal variability of the Amazon River Basin Water Budget, Climate Dynamics, 24, 11-22, 2005.

Nesbitt, S. W. and Zipser, E. J.: The diurnal cycle of rainfall and convective intensity according to three years of TRMM measurements, J. Climate, 16, 1456-1475, 2003.

Savenije, H. H.G.: The importance of interception and why we should delete the term evapotranspiration from our vocabulary, Hydrological Processes, 18, 1507-1511, 2004.

Savenije, H. H. G.: The Runoff Coefficient as the Key to Moisture Recycling, Journal of Hydrology, 176, 219-225, 1996a.

Savenije, H. H. G.: Does moisture feedback affect rainfall significantly?, Physics and Chemistry of the Earth, 20, 507-513, 1996b.

Savenije, H. H. G.: New definitions for moisture recycling and the relation with land-use changes in the Sahel, Journal of Hydrology, 167, 57-78, 1995.