All the arguments presented in this article are based on Global Climate Model (GCM) results that fail to reproduce the current climate and particularly the hydrological cycle, which is fundamental to the existence of the Amazon Biome. In the maps shown in the article, the reader gets the impression that the variation of temperatures and rainfall presented is a “continuum,” both in space and time. However, this is not real. A GCM is a very complex computer program with millions of lines of codes that relies on a supercomputer to solve a given working hypothesis or sensitivity test, in this case, the impact of Amazon deforestation on the global climate. However, the computer does not know what a continuum is. Input data for GCM must be discretized. That is, data assimilation in the GCM is done at grid points both horizontally and at altitude (three-dimensional grid), as shown in Figure 1.
The Geophysical Fluid Dynamics Laboratory (GFDL) GCM has a horizontal spacing between grid points, i.e., a spatial resolution of 1° latitude by 1° longitude [111km x 111km] and is integrated with a 30-minute time step. Data of weather variables (temperature, humidity, wind, etc.) for initializing the simulation are reported to the GCM at each grid point, 111km apart. This means that the GCM has no “information” or data about what happens between one point and another. All spatial scale processes less than the distance between grid points [sub-grid processes] such as hydrological cycle, the vegetation [albedo, evapotranspiration, and CO2 emission by forest], soil properties, topography, must be “parameterized.”
In particular, physical processes of vertical air turbulence, convection (upward movement of humid air) required for cloud formation and rain are scale processes of the order of 100m to 1,000m and are not “perceived” by the model. For example, a cumulonimbus cloud, 20 km in diameter and 12 km high, located between grid points, can produce rainfall of up to 50mm locally, which would not be simulated by the model as it has no information about the presence of this cloud. Simple, empirical or observation-based mathematical formulas are created in an attempt to reproduce the actual sub-grid physical processes. Therefore, the parametric equations used in GCM codes are only approximations of the actual physical processes that occur in the climate system. Some of them can be quite representative, but others may be crude because the physical processes they represent are either not well understood (such as evapotranspiration) or too complex to be included in code due to computational constraints. Among the parameterizations, the formation of clouds and rain (hydrological cycle) deserves mention as it remains one of the most difficult challenges, if not the greatest, responsible for the large difference of results among GCM.
According to the Nyquist Sampling Theorem, the distance between grid points is a filter for physical processes less than twice the distance between grid points. In the case of the GFDL’s GCM, all physical processes with a spatial scale of less than 222 km in the tropics are filtered and not “perceived” by the model. Besides, these parameterizations have to be ‘calibrated’ or ‘tuned’ for the model results to approach the observed reality. In this respect, there are two points worth noting.
First, most of GCMs were “calibrated” with data observed from 1975 to 2000, a period when, admittedly, the global climate was warm. Therefore, this practice is biased as it “tunes” GCMs in the warm phase of the natural internal variability of the climate system and makes GCMs overly sensitive to CO2 variations. Second, GCMs are specifically designed to respond to the increased concentration of CO2 in the atmosphere, so that, according to Intergovernmental Panel on Climate Change (IPCC) reports, if CO2 concentration remains fixed, GCMs show no significant global warming, some even show cooling. Scenarios of future CO2 concentrations (SSP and RCP) are fictitious, created by the human mind, and some of them are impossible to realize, as it is the case of RCP8.5, which predicts a CO2 concentration 3 times higher than the current one for 2100. In short, there is a fierce debate in the scientific community as to the accuracy of GCM results in sensitivity testing and concerning usefulness to predict future climates. Therefore, all GCM/GFDL simulation results presented on Amazon deforestation, including remote impacts on the global climate, such as warming in the American Midwest and the Arctic, are highly questionable.
An example of GCM’s unreliability in predicting the future climate is shown in Figure 2, in which a) shows forecast of rainfall anomalies for March 2019 with the GCM/GFDL initialized with October 2018 data, and (b) March 2019 observed anomalies, with data from the Global Precipitation Climatology Center (GPCC/ESRL/NOAA). The error of the GCM/GFDL forecast, particularly over the Amazon, made only 5 months in advance is evident. If the GCM / GFDL grossly errs 5 months in advance, how reliable can be its forecasts for the years 2050, 2100, and how useful are these forecasts for planning activities in agriculture and hydropower generation?
Curiously, GCM/GFDL has a total of about 64,000 surface grid points, of which only about 500 [0.8%] are in the Amazon. It is intriguing that only 0.8% of grid points should have a noticeable influence on global temperatures and rainfall in model simulations.
The statement that the evapotranspiration of the forest is what generates the water vapor transported inland in the so-called flying rivers, and that transforming the forest into pasture would reduce rainfall over Brazil by 25% is only a result of GCM/GFDL, without any verification or finding. The Amazon is not essential for the distribution of rainfall over other remote regions of South America because the Amazon region is not a source of moisture to the atmosphere. Regarding the climate, the Amazon has a stable water balance. The main source of humidity for Amazon rains is the North Atlantic Ocean, mainly during the Southern Hemisphere summer. Data observed between 1999-2014 suggest that, on average and in round numbers, enters the region the equivalent of 500,000 m3/s of humidity, of which 80% is transformed into rain, and the remaining 100,000 m3 / s “passes straight” over the region. Half of the 400,000 m3 / s of rainfall falling into the basin leaves it through the Amazon River [200,000 m3 / s], and the other half is recycled by evapotranspiration and incorporated into the moisture stream that reaches other regions of South America. In other words, 300,000 m3 / s, out of the 500,000 m3 / s [60%] originally from Atlantic evaporation, reach other regions outside of Amazonia, with the remaining 200,000 m3 / s being returned to the Atlantic by the river discharge. A tree or forest is not a “water-making machine” , but it only recycles rainwater previously stored in the soil. Although there is a forest-atmosphere interaction, in the long run, the forest exists because it rains and not the other way around. It is proved by reductio ad absurdum that, if the forest were a source of moisture, the region would have become a desert since it stabilized after the end of the last ice age, about 15,000 years ago.
The key geophysical element for directing moisture flow from the Atlantic to other regions of South America is the formidable barrier that the Andes Cordillera raises to moisture flow. Another element is a direct atmospheric circulation cell known as the Hadley-Walker Cell, which normally forms and will always form, as the Sun inevitably warms the surface of the South American continent during the southern summer. As a result, the air becomes less dense and rises (convection), carrying moisture and producing clouds and rain. The forest will very likely interact with the atmosphere to intensify this circulation cell in regular years. Years in which this cell does not form are an exception. For example, it has been observed that when there is a strong El Niño event such as the 2014-2016 event, this circulation cell is inhibited and the Amazon Basin experiences a severe drought. This would not happen if the forest were the main cause of the existence of this atmospheric circulation cell. However, the warming of the surface by the Sun, and the consequent circulation cell, will always exist regardless of the existence of the forest.
The carbon stock in the Amazon is about 70 to 80 billion tons of carbon (GtC) – assuming the dry biomass density is between 250 and 300 tC per hectare across the biome – which, if fully released into the atmosphere, could theoretically increase the global atmospheric CO2 concentration by about 35 ppmv. However, a simple calculation shows that at the current rate of deforestation of 7,500 km2 per year, the total release of this stock would take about 700 years to be completed, assuming zero regional carbon addition, that is, no regrowth during this period. On the other hand, measurements made in Central Amazonia in 1987 during the Atmospheric Boundary Layer Experiment (NASA/INPE) ABLE-2B revealed photosynthetic assimilation of 4.4 kilograms of carbon per hectare per hour [kgC / ha / hour] during the day and a respiration loss of 2.57 kgC / ha / hour during the night. Assuming that these numbers could be generalized to the 550 million hectares of the Amazon Biome, there would be 9 GtC / year of carbon assimilation by photosynthesis. Considering that global human activities emit about 10 GtC/year today, this assimilation corresponds to practically 100% of current carbon emissions. If one accepts the absurd hypothesis advocated by the IPCC that CO2 is the major controller of the global climate, then the Amazonian vegetation cover must be preserved.
The media claim that the destruction of the Amazon biome is accelerating. However, Figure 3 shows that the rate of deforestation has been much higher in the past, peaking in 1995 [29,059 km2] and between 2002 and 2005, with a secondary peak in 2004 [27,772 km2]. In 2018, the estimated rate was 7,900 km2, a 72% reduction from 1995. In the case of Brazil, it is also necessary to distinguish the “Legal Amazon” – which is a territory of 5.2 million km2 demarcated for tax incentives – from the Amazon Biome that covers about 65% of the Legal Amazon. The parts of the Legal Amazon that suffer from anthropic pressure are the southern and eastern regions, characterized by transitional biomes such as cerradão and cerrado, which are not covered with tropical rain rainforests. Note that 84% (or more) of the Amazon Biome in Brazilian Territory is preserved according to data from the Ministry of the Environment, an area equivalent to the territories of Germany, Spain, Finland, France, the Netherlands, Ireland, Norway, Portugal, United Kingdom, and Sweden combined.
Let me make it perfectly clear that I am not supporting widespread deforestation, but only expressing the rational and most likely intuitive physical phenomenology. The Amazon Biome encompasses 5.5 million km2, while the surface of Planet Earth has 510 million km2, and its oceans cover 361 million km2. Therefore, the Amazon Biome corresponds to 1% of the surface of Planet Earth and 1.5% of its oceans. In principle, deforestation of 50% or 100% of the Amazon would not affect the global climate because it is a region of small proportions compared to the oceans area (71%), these being one of the major global climate controllers.
Furthermore, if widespread deforestation should occur, the surface of the region would become aerodynamically less rough and, intuitively, one would expect winds to intensify at lower atmospheric levels where the highest moisture concentrations are found, carrying more moisture out of the basin and increasing the availability of moisture to be transformed into rain in regions to the south of the Amazon. In fact, impacts on the global climate and global CO2 concentration [Paris Climate Agreement, 2015] are not arguments for maintaining the forest. The main arguments are the conservation of biodiversity and the protection of soils, preventing erosion, siltation of river beds, change in the quality of waters, and all life that depends on them.
Considering that about 25 million people live in the region, most of them under a living standard incompatible with human dignity, the solution is to identify and test innovative methods of regional development that exploit natural, renewable and non-renewable resources, conserving the vegetation cover.