A bathtub is being filled [sun], slightly faster than it is being drained [heat radiation]. You have a few floats, measuring the depth of the water. The depth would go up smoothly, except there’s a kid splashing around in the bath.

Sometimes the kid lies back in the water, in which case the overall water level goes up [El Nino], but with waves, so that some floats go down.
Sometimes the kid sits up, in which case the overall water level temporarily goes down [La Nina], but with waves, so a few of the floats go up.

The kid splashes around the whole time, jiggling all floats second by second.

At any point in time, there is a certain amount of water, but the average as measured by 1% of the floats is subject to lots of jiggles.

Still, the water *is* going up, as long as more as coming in than draining out, and the physics of GHGs say that we’re slowly plugging the drain.

The post Bathtub Analogy: Why the Global Temperature Giggles appeared first on Atmoz.

What is a monsoon?

The word monsoon derives its name from the Arabic word ’season’. Sailers in the Indian Ocean would use the word to describe the change in wind direction throughout the year. They had noticed that during the summer the winds blow from a different direction than during the winter. The word monsoon refers to the seasonal reversal of winds, and not to changes in the precipitation (National Weather Service).

The most famous monsoon is the Indian Monsoon, which not surprisingly occurs over the Indian subcontinent. The formation of a monsoon is similar to the formation of the land- or sea-breeze. Because of the difference in heat capacities, the land surface heats faster than the ocean during the day. This causes the formation of a thermal low. The resulting surface pressure difference between the land and ocean cause the wind to blow from the ocean to the land (during the day), or from the land to the ocean (during the night).

The monsoon is formed on the same principles as the sea-breeze, except over larger areas and longer times. Instead of a diurnal variation in wind direction as occurs with the sea-breeze, the changes in the monsoon winds occur on seasonal time scales. During the summer months, the Mojave Desert becomes extremely hot, and a semi-persistent thermal low develops over the region. This is what causes the winds to change directions from northeasterly in the winter months to southwesterly in the summer months.

So What is the Monsoon?

If we already know what the monsoon is and why it forms, then what’s the big idea? Part of the problem is in the predictability of its onset and duration. Because roughly one-half of the precipitation in the Southwest occurs during the monsoon, there are important implications of changes in the monsoon due to climate change. With the increase in temperatures, is the desert Southwest going to turn into a dust-bowl or will the changes be more subtle and complicated (see e.g. Seager, 2007).

Kursinski et al. use the Global Positioning System (GPS) satellites to determine the amount of precipitable water vapor (PWV) in the atmosphere. The main purpose of the paper is to describe the use of the GPS technique to allow high temporal resolution water vapor measurements to help determine the source of PWV in an effort to increase the skill in predicting summer precipitation.

They introduce an innovative approach to defining the start of the monsoon. The National Weather Service defines the start date of the monsoon as “when the average daily dewpoint is 54 degrees or greater for 3 consecutive days.” However, since the monsoon describes a seasonal shift in winds, this is not the best approach. Kursinski et al. “define the monsoon as beginning during the 5 to 7 day onset period commencing on July 1 over which PWV values grew to sustained high values and the PWV-weighted winds at Empalme shifted from westerly to southeasterly and remained there.” This definition has one major advantage to the NWS onset date: it incorporates the direction of the wind.

The analysis of Kursinski et al. show that during the 2004 North American Monsoon there were two distinct phases which they label as “sub-synoptic scale “and “synoptic scale”. Each of these phases has different statistical characteristics. For instance, during the sub-synoptic scale phase there is increased moisture variability at scales smaller than 90 kilometers. They attribute this variation to changes in moist convection and mountain circulation.

During the synoptic scale phase, the statistical nature of the monsoon changes. The variations are no longer on small scales, but on the large scales. They postulate that this could be due to the advection of moist air into (or out of) the North American Monsoon region. One major question unable to be answered by this study is whether the two-regime nature of the monsoon is typical of the North American Monsoon or whether it was simply a one-year occurance.

Implications

The North American Monsoon system is not well researched, and not well understood. While the basic physical principles for the existance of the monsoon have been known for some time, the inability to predict the onset and duration of the monsoon point to significant gaps in the scientific understanding of an important precipitation source for the Southwest United States. With the likely northward progression of mid-latitude winter storms due to global warming, it is important to fully understand the consequences of changes in the monsoon.

References:

Kursinski, E.R., Bennett, R.A., Gochis, D., Gutman, S.I., Holub, K.L., Mastaler, R., Minjarez Sosa, C., Minjarez Sosa, I., van Hove, T. (2008). Water vapor and surface observations in northwestern Mexico during the 2004 NAME Enhanced Observing Period. Geophysical Research Letters, 35(3) DOI: 10.1029/2007GL031404

National Weather Service (NWS), Flagstaff Forecast Office, The Monsoon.

Seager, R., et al. (2007), Model predictions of an imminent transition to a more arid climate in Southwestern North America, Science, 316, 1181–1184, doi:10.1126/science.1139601.

The post Water Vapor Measurements during the North American Monsoon appeared first on Atmoz.

Scripps Institution of Oceanography had several monitoring sites scattered throughout the World, mostly on remote islands in the Pacific. The Mauna Loa record is just the one that extends back furthest in time. We can see that all of the observations show excellent agreement.

There is something remarkable about that graph: no matter where the measurements were taken, the CO2 concentration is always about the same. There are squiggles due to the annual cycle that aren’t the same, but that’s because of the location of the sites. The Antarctica station, (SPO – cyan), has almost no annual cycle because of its remote location.

The post CO2 is Still Rising – Even at Locations other than Mauna Loa appeared first on Atmoz.

For this model I will assume that the ocean consists of two parts: the surface ocean and the deep ocean. The surface ocean is uniform in depth, temperature, and salinity. The depth of the surface ocean is 500 meters. The average initial temperature of the upper ocean is 14C. The deep ocean is everything else, and is assumed to not change.

The volume of water in the ocean is given by the equation: V=A*d, where A is the surface area of the ocean and d is the depth of the ocean. We also know that the mass of an object is equal to its volume multiplied by its density; m=V*ρ. We can solve these equations for d, the depth of the ocean.

d = ^m/(_ρ*A)

And since we’re interested in the change in d, or Δd, that’s equal to

Δd=d-d0, where d0 is the initial height of the ocean, 500m.

In our equation above, we have the change in depth as a function of density, and we’re assuming that the mass of the ocean and its surface area do not change. However, that’s not really interesting, so let’s find how the density of sea water changes with temperature. As stated above, the salinity, or the saltiness, of the ocean will be held constant. So the density is only dependent upon the temperature. We could go through the calculations here to figure out the density as a function of temperature, or we could cheat and use this handy water density calculator. I’ve reproduced some values in the table below:

We now have all we need to figure out how much sea levels will rise do to thermal expansion.

This simplistic model of the surface ocean shows that if the average sea surface temperature rises by 1.4C, sea levels will rise by about 6 inches.

The post Sea Level Rise due to Thermal Expansion appeared first on Atmoz.

The accompanying Fig 1 shows a comparison of our data with the official NOAA data taken at Tucson International airport. From 1993 to 1996, the data was from our old site at the Medical School, taken with a non-aspirated temperature sensor. In May 0f 1997 the site was changed to its current location but data was still being taken with a non-aspirated sensor. The aspirated RM Young instrument was installed early in 1998 and has been in use since then. The installation of this new system is responsible for the drop in U of A temperatures relative to the NOAA temperatures since 1998. Indeed, non-aspirated instruments can read 4-5 degrees warmer than aspirated instruments during our hot sunny days.

The second figure presents our temperature data back to 1945 and also data from the NOAA airport station from 1948. I think it is quite clear that the 2 records follow each other very closely. Both show a definite warming trend. How much of this is due to the growth of Tucson and how much can be attributed to other causes is impossible to determine at this time, but the trend is clearly present in both data sets. [Emphasis mine.]

There are no “official” photos of the Tucson ASOS on the NCDC station locator, but I’ve managed to round up a couple. The first is from NOAA Surface Observations Program, taken in August 1996. The second is from NOAA Western Region Notes [PDF, top of page 3], taken sometime in 2005. Both photos indicate this is a “good” station.

The post More on Tucson USHCN appeared first on Atmoz.

Tropical cyclones form only under specific environmental conditions. Anthropogenic climate change might alter the geographical areas where tropical cyclones can develop. Using an ensemble of regional climate models, we find an increase in the extremes of cyclone intensity over the Mediterranean Sea under a climate change scenario. At least for the most sensitive model, the increase in intensity is clearly associated with the formation of tropical cyclones. Previous studies did not find evidence of changes in the projected areas of formation of tropical cyclones (Intergovernmental Panel on Climate Change, 2007; Walsh, 2004; Lionello et al., 2002). Those studies were based either on relatively low-resolution global climate models or on one particular regional climate model. The use of a multi-model ensemble of relatively high-resolution regional climate models has allowed us to detect for the first time a risk of tropical cyclone development over the Mediterranean Sea under future climate change conditions.

I am skepical that there will ever be any hurricanes in the Mediterranean Sea. All known tropical cyclones have developed in the tropics, where the mean flow is from east to west. The Mediterranean region, on the otherhand, has mean westerly mean flows, even in the summer months when potential tropical storms may develop. There is nothing that says a tropical storm needs to move from east to west, but since that is the mean flow in the tropics, that’s what happens. When the storms move to far north or south, they either get absorbed by the mean westerly flow or get intensified into a subtropical storm.

The Mediterranean Sea has it’s highest sea surface temperatures in the eastern region, near Israel and Egypt. Even now, these temperature are high enough, for a brief period of time during the summer, for tropical storm development. In global warming scenarios, the sea surface temperature of the Mediterranean will increase, and according to the authors, possibly result in tropical storm formation.

However, as I mentioned above, the mean flow is from west to east. The Western Mediterranean is very cool; it is not warm enough for tropical storm development. Even assuming a warming of 2C, this will not be warm enough. The only place for tropical storm development will be in the eastern part of the sea. However, since the mean flow is from west to east, there will probably not be enough time for development into hurricanes. Since these storm will not be in the tropics, they will technically be sub-tropical depressions and sub-tropical storms. It is my opinion that even under global warming scenarios sea, the meteorological conditions will not be favorable for hurricane development in the Mediterranean Sea.

The post Tropical cyclones over the Mediterranean Sea in climate change simulations appeared first on Atmoz.

I don’t actually recommend reading this. But one gem that they propose is that there isn’t a thing called an average temperature. Of course there is. When attempting to derive such a temperature, Gerlich and Tscheuschner arrive at a value of 87.6 C. This is clearly wrong. If the Earth were that warm, humans wouldn’t exist. They then “explain” how climatologists get their value to explain the greenhouse effect as follows:

This fictitious [greenhouse] effect is based on the assumption that one should have an average effective temperature of -18 [degrees] C. One will get this if one weights the solar constant with a factor of 0.7 and inserts a quarter of the solar constant into the “radiative balance” equation. The factor of a quarter is introduced by “distributing” the incoming solar radiation seeing a cross section σ_Earth over the global surface ΩEarth.

[Added August 7, 2007: This post has been linked from a comment in a Scienceblogs.com post. For those who don’t believe in the greenhouse effect, please explain how the average temperature on the moon is lower, in spite of the fact that it has a lower albedo.]

Actually, the value of -18 C falls right out of the equations, not the other way around. Assume that the sun radiates at a certain temperature such that it can reasonably be modeled by the blackbody curve for some effective temperature. This shouldn’t be that hard to do, even Gerlich and Tscheuschner do so in their paper. By the time this radiation reaches the Earth, it’s intensity has decreased according to the ¹/_R² law. Again, this is exactly what Gerlich and Tscheuschner do. At the Earth, this value is roughly constant – 1367 ^W/_m². Gerlich and Tscheuschner do not use this number, they keep their equation in terms of the temperature of the sun, radius of the sun, and distance from the Earth to the sun. It doesn’t matter, the finals answers will end up the same.

Therefore, the total energy absorbed by the Earth is related to its albedo and its radius.

E_A = (1-A)S₀πR²

The term 1-A is the percentage of incoming solar radiation absorbed by the Earth; the albedo (A) is the percentage reflected. S₀ is the solar constant. And πR² is the cross-sectional area of the Earth that absorbs radiation. The dark side of the Earth cannot absorb radiation from the sun.

The total energy emitted by the Earth is related to its temperature and its radius.

E_E = σT⁴4πR²

Because the Earth emits radiation from its entire surface and not just the side facing the sun, the surface area of the Earth is used (4πR²) is used instead of the cross-sectional area. The σT⁴ term is the blackbody emission for an object at a given temperature.

Setting the two equations equal – assuming the energy absorbed equals the energy emitted – and simplifying, we see that

(1-A)S₀ = 4σT⁴

So, for a given A and S₀, we can find the effective temperature. In the case of the Earth, the albedo (A) is about 0.3, so 1-A is 0.7, which magically explains where that factor comes from that Gerlich and Tscheuschner couldn’t explain. The factor of 4 is just a consequence of the fact that the Earth can only absorb radiation on the side facing the sun, but emits in all directions. When the values are plugged in, we (and Gerlich and Tscheuschner) get a value of -18 C.

So, Gerlich and Tscheuschner couldn’t figure out where the magical values of 0.7 and 0.25 [1/4] came from, but they are just misleading their readers. They [should] know how to compute an effective temperature. And they [should] know that such a value exists, and is physically meaningful.

The post Falsification Of The Atmospheric CO2 Greenhouse Effects appeared first on Atmoz.

Abstract: Through three-dimensional large-eddy simulations of marine stratocumulus we explore the factors that control the cloud spectral relative dispersion (ratio of cloud droplet spectral width to the mean radius of the distribution) as a function of aerosol number concentration and the extent to which the relative dispersion either enhances or mitigates the Twomey effect. We find that relative dispersion decreases with increasing aerosol number concentration (for aerosol number concentrations less than about 1000 cm^−3) because smaller droplets resulting from higher aerosol number concentrations inhibit precipitation and lead to (1) less spectral broadening by suppressed collision and coalescence processes and (2) more spectral narrowing by droplet condensational growth at higher updraft velocity because reduced drizzle latent heating at cloud top results in increased boundary layer turbulent kinetic energy production by buoyancy and thereby stronger turbulence. Increased spectral broadening owing to increased cloud-top entrainment mixing, also as a result of increased boundary layer turbulence, is relatively insignificant compared with outcomes 1 and 2. The coefficient k, an important parameter that relates cloud droplet effective radius and volume mean radius in large-scale models, is a function of skewness and relative dispersion of the distribution and is negatively correlated with relative dispersion. Increasing k with increasing aerosol number concentration leads to maximum enhancement of the cloud susceptibility (the change of cloud optical depth due to change of cloud droplet number concentration) over that attributable to the Twomey effect alone by about 4.2% and 39% for simulated FIRE and ASTEX cases, respectively.

As with the last paper I talked about in this series, this one is again a modeling study. The good news is that it is by the same authors as the last one, and uses the same model. That was good because I didn’t need to wade through all the modeling talk, that I only vaguely understand. As the title of this paper says, it is about the effect of aerosol number concentration on cloud droplet dispersion. Remember that relative dispersion is just a measure of the width of the size distribution of the liquid water droplets. They attempt to show the factors that control dispersion between clean and polluted conditions, and to what extent this change has on the Twomey effect.

The paper presents a flow chart representing the physical mechanisms that result from an increase in N_a, the aerosol concentration and potential cloud condensation nuclei (CCN). An increase in N_a results in more numerous, smaller droplets. This is because there are more CCN for the water to condense onto. This in turn leads to suppressed drizzle (Albrecht, 1989). Less drizzle means that there are fewer large droplets that fall into and collect the smaller droplets, called collision and coalescence. This leads to a smaller relative dispersion. Suppressed drizzle also leads to less latent heat caused by the condensation of water at cloud top, leading to more turbulent kinetic energy (TKE) production. This leads to stronger updrafts, and condensational spectral narrowing, or a smaller relative dispersion. Increased TKE production also implies a larger entrainment mixing, which leads to a larger relative dispersion, but they note that the effect is small.

The results of the model experiment is that an increase in N_a does in fact lead to a decrease in the relative dispersion. They note that this is consistent with several observations (Pruppacher and Klett, 1997; Miles et al., 2000; Yum and Hudson, 2005). However, they also point out that other observations noticed the exact opposite effect, an increase in N_a leads to an increase in the relative dispersion (Martin et al., 1994, Ackerman et al., 2000; McFarquhar and Heymsfield, 2001; Liu and Daum, 2002).

As can be seen, there is no general agreement in the literature what effect an increase in aerosol has on the spectral dispersion of the liquid water droplets. There seems to be an equal number of studies that find opposite results. Part of this could be contributed to experimental design, as the paper points out. Most of the past studies of this nature were done within the size range of the Forward Scattering Spectrometer Probe (FSSP) instrument. As a result, the relative dispersion was only calculated over the cloud droplet spectrum, and did not include the drizzle. The addition of the drizzle tends to increase the relative dispersion, according to Lu and Seinfeld. This was a result from their model, and as far as I know, has not been shown in field studies.

References:
Ackerman, A. S., O. B. Toon, J. P. Taylor, D. W. Johnson, P. V. Hobbs, and R. J. Ferek, Effects of aerosols on cloud albedo: Evaluation of Twomey’s parameterization of cloud susceptibility using measurements of ship tracks, J. Atmos. Sci., 57, 2684–2695, 2000.
Albrecht, B., Aerosols, cloud microphysics, and fractional cloudiness, Science, 245, 1227-1230, 1989.
Liu, Y. G., and P. H. Daum, Anthropogenic aerosols: Indirect warming effect from dispersion forcing, Nature, 419, 580–581, 2002.
Martin, G. M., D. W. Johnson, and A. Spice, The measurement and parameterization of effective radius of droplets in warm stratocumulus clouds, J. Atmos. Sci., 51, 1823–1842, 1994.
McFarquhar, G. M., and A. J. Heymsfield, Parameterizations of INDOEX microphysical measurements and calculations of cloud susceptibility: Applications for climate studies, J. Geophys. Res., 106(D22), 28, 675–28, 698, 2001.
Miles, N. L., J. Verlinde, and E. E. Clothiaux, Cloud droplet size distributions in low-level stratiform clouds, J. Atmos. Sci., 57, 295–311, 2000.
Pruppacher, H. R., and J. D. Klett, Microphysics of Clouds and Precipitation, 976 pp., Springer, New York, 1997.
Yum, S. S., and J. G. Hudson, Adiabatic predictions and observations of cloud droplet spectral broadness, Atmos. Res., 73, 203–223, 2005.

The post Aerosol Number Concentration and Cloud Droplet Dispersion appeared first on Atmoz.

Lu and Seinfeld in the Journal of the Atmospheric Sciences, Study of the Aerosol Indirect Effect by Large-Eddy Simulation of Marine Stratocumulus [PDF].

Abbreviated Abstract: A total of 98 three-dimensional large-eddy simulations (LESs) of marine stratocumulus clouds covering both nighttime and daytime conditions were performed to explore the response of cloud optical depth (Ï„) to various aerosol number concentrations (Na = 50–2500 cm−3) and the covarying meteorological conditions (large-scale divergence rate and SST)… The second indirect effect is found to enhance (reduce) the overall aerosol indirect effect for heavily (lightly) drizzling clouds; that is, Ï„ is larger (smaller) for the same relative change in Na than considering the Twomey (first indirect) effect alone. The aerosol indirect effect (on Ï„) is lessened in the daytime afternoon conditions and is dominated by the Twomey effect; however, the effect in the early morning is close but slightly smaller than that in the nocturnal run.

As usual, since this is a modeling study, I will gloss over the specifics of the model. Again, not because they aren’t important, but because I don’t fully understand them. This model is RAMS, the Regional Atmospheric Modeling System. Lu and Seinfeld wrote, “In the basic droplet activation scheme implemented in the RAMS modle the aerosols are assumed to have a constant size distribution is space and time.” This is describing the non-liquid water aerosols, or the potential CCN particles. While this is technically untrue, the size distribution of CNN changes with respect to space and time, I am unsure what the effects this would have on the outcome of this particular study. In fact, it’s probably a good thing to keep the CCN distribution constant to isolate the effects of the other changed variables. But I would find it interesting to see the effects of the changes in CCN sizes in both space and time.

The results reported seem to be contradictory. In the beginning of the second paragraph in the results section it is stated that “cloud LWP varies only slightly with aerosol concentration, particularly at high aerosol loadings. There is a slight decrease of LWP as N_a increases from 50 to 500 cm^-3 in each of the cases. The effect of high pollution loadings in suppressing drizzle formation (Albrecht 1989) is evidenced as precipitation disappears when N_a exceeds 500 cm^-3.” Later, at the end of that same paragraph is the following “the result shows cloud LWP is slightly higher for the clean clouds in contrast to the typical secondary aerosol effect.” How do they resolve this? Not very well in my opinion, if at all. From their figures, it appears that at aerosol loading under 500 cm^-3 (clean clouds), there is an accumulation of precipitation at the surface. And that at aerosol loadings of 500 cm^-3 or greater (dirty clouds), there is no accumulation of precipitation. Thus, showing the traditional second indirect effect. However, in the figure right next to that one, for aerosol loadings less than 500 cm^-3, the liquid water path (LWP) is greater than for aerosol loadings equal or greater than 500 cm^-3. To first order, it doesn’t make sense why precipitation should decrease and LWP decrease at the same time. Though there are at least two possible solutions. First, there could be more precipitation from the cloud, but that it is being evaporated below the cloud deck. Also, the favorite cause of entrainment; dry air from above cloud top can be mixed with the moist air to lower LWP. Both hypotheses are posited, but its not articulated which one is the primary cause, if any.

The effective radius of the cloud size distribution, r_e, is defined as the third moment of the size distribution divided by the second moment. The spectral shape parameter, k, is inversely proportional to effective radius. But, effective radius is proportional to the LWP. Therefore, it is important to properly understand the effects of aerosol loading on LWP because it has implications for the calculations of spectral dispersion (shape).

The post Aerosol Indirect Effect in Marine Stratocumulus appeared first on Atmoz.

Abstract: Parameterization of effective radius (r_e) as proportional to the cube root of the ratio of cloud liquid water content (L) to droplet concentration (N), i.e., r_e=α(L/N)1/3, is becoming widely accepted. The principal distinction between different parameterization schemes lies in the specification of the prefactor α. This work focuses on the dependence of α on the spectral dispersion of the cloud droplet size distribution. Relationships by Pontikis and Hicks [1992] and by Liu and Hallet [1997] that account for the dependence of α on the spectral dispersion are compared to each other and to cloud microphysical data collected during two recent field studies. The expression of Liu and Hallet describes the spectral dependence of α (or r_e) more accurately than the Pontikis and Hicks relation over the observed range of spectral dispersions. The comparison shows that the different treatments of α as a function of spectral dispersion alone can result in substantial differences in r_e estimated from different parameterization schemes, suggesting that accurately representing re in climate models requires predicting α in addition to L and N.

The relative dispersion of the cloud is the ratio of the standard deviation of the distribution to the mean radius. A low dispersion means that the particles are all around the same size; a high dispersion means the sizes are more spread out. As mentioned in the last article, spectral dispersion is thought to play an important role in the anthropogenic modification of clouds.

Cloud droplet effective radius, r_e, is a key variable that is used in the radiative transfer calculations of liquid water clouds. In some global climate models, r_e is parameterized as a function of liquid water content and cloud droplet number concentration, such that r_e is directly proportional to L^1/3 and inversely proportional to N^1/3. r_e is also directly proportional to α, the prefactor. If all the cloud droplets are of the same size, monodisperse, then the prefactor is about 62. This is obviously non-physical because no cloud can have a monodisperse size distribution. For polydisperse size distributions, the prefactor is increased; the broader the size distribution, the larger the prefactor.

This paper talks about several of the parameterizations of the prefactor, and how the effective radii derived from using it compares with field data. The best model is that developed by Liu et.al. (1995). The cynic in me notes that it was written by the same author as this paper. At large spectral dispersions (greater then 1.0), all but the Liu parameterization underestimate the effective radius of the distribution. The Liu parameterization actually overestimates the effective radius at high spectral dispersions. The parameterizations all did fairly equal at very low spectral dispersions (less than 0.4). The paper notes that the bias introduced by the other parameterizations is enough to cause problems in climate models.

The main conclusion of this paper is that it demonstrates that it is necessary in global climate models to include the spectral dispersion in the calculation of effective radius. Just a short summary of this paper because I’m not a modeler. But it does show that including the effects of spectral dispersion are important to be included in any climate change model.

References:
Liu, Y., L. You, W. Yang, and F. Liu, On the size distribution of cloud droplets, Atmos. Res., 35, 201-216, 1995.

The post Spectral Dispersion and Cloud Droplet Effective Radius appeared first on Atmoz.