Possible climate interaction mechanism between terrestrial climatic parameters and solar activity

Possible climate interaction mechanismbetween terrestrial climatic parameters and

solar activity

J. Osorio R.∗

Instituto de GeofısicaUniversidad Nacional Autonoma de Mexico

Mexico D.F., C.P. 04510.

[email protected]

B. Mendoza O.Instituto de Geofısica

Universidad Nacional Autonoma de MexicoMexico D.F., C.P. 04510.

[email protected]

February 22, 2012

Abstract

The solar activity has been proposed as one of the main factors of Earths climatevariability, however biological processes have been also proposed. The Dimethyl-sulfide (DMS) is the main biogenic sulfur compound in the atmosphere. The DMSis mainly produced by the marine biosphere and plays an important role in theatmospheric sulfur cycle. Currently it is accepted that terrestrial biota adapts andinfluences to environmental conditions through regulations on the chemical compo-sition of the atmosphere In the present study we used different methods of analysisto investigate the relationship between the DMS, Low Clouds, Ultraviolet Radia-tion A (UVA) and Sea Surface Temperature (SST) in the Southern Hemisphere. Wefound that the series analyzed have different periodicities which can be associatedwith climatic and solar phenomena such as ENSO, The Quasi-Biennial Oscillationand changes in solar activity mainly in the cycle of the 11 years. Also, we found ananti-correlation between DMS and UVA, the relation between DMS and low cloudsis mainly non-linear and there is a correlation between DMS and SST. The resultssuggest a positive feedback interaction among DMS, solar radiation and low cloudsat time-scales shorter (2 years) than the solar cycle.

Keywords: Dimethylsulfide, Solar Activity, Sun-Earth Relations, Wavelet Analysis,Vector Autoregressive Analysis.

∗Earth Sciences Graduate Programme - UNAM

J. Osorio & B. Mendoza (2012), 1

1 Introduction

The solar activity has been proposed as an external factor of Earth’s climate change, solarphenomena such as total and spectral solar irradiance could change the Earth’s radiationbalance and hence climate (Solanki, 2002 ). However, biological processes have also beenproposed as another factor of climate change through its impact on cloud albedo. Oneof the most important issues regarding the Earth function system is whether the biotain the ocean responds to changes in climate (Charlson et al., 1987, Miller et al., 2003,Sarmiento et al., 2004 ). According to several authors, the major source of cloud conden-sation nuclei (CCN) over the oceans is dimethylsulfide (DMS) (e.g., Andreae and Crutzen,1997; Vallina et al., 2007 ).

Dimethylsulponiopropionate (DMSP) in phytoplankton cells is released into the watercolumn where it is transformed into dimethylsulfide. The DMS diffuses through the seasurface to the atmosphere, where it oxidizes to form a range of products, among themSO2. This compound oxidizes to from H2SO4, which forms sulphate particles that actas CCN. The DMS concentration is controlled by the phytoplankton biomass and by aweb of ecological and biogeochemical processes driving by the geophysical context (Simo,2001 ). Solar radiation is the primary driving mechanism of the geophysical context andis responsible for the growth of the phytoplankton communities. Clouds have a major im-pact on the amount of heat and radiation budget of the atmosphere. Clouds modify bothalbedo (short-wave) and long-wave radiation. In particular, for low clouds over oceans,the albedo effect is the most important result of cloud radiation interaction and has a netcooling effect on the climate(Chen et al., 1999; Rossow, 1999 ).

The DMS, solar radiation and cloud albedo are hypothesized to have a feedback interac-tion (Charlson et al., 1987; Shaw et al., 1998; Gunson et al., 2006 ). This feedback canbe either negative or positive. A negative feedback process requires a positive correlationbetween solar irradiance and DMS: increases in solar irradiance reaching the sea surfaceincrease the DMS, augmenting the CCN and the albedo. An overall increase in the albedoproduces a decrease in the irradiance reaching the sea surface and thus cooling occurs. Apositive feedback requires an anti-correlation between solar irradiance and the DMS: in-creases in solar irradiance produce a decrease of the DMS, the CCN and the albedo; a netreduced albedo allows more radiation to be absorbed, producing heating. Using seasonaland annual time scales, a first attempt to find correlations between a global database ofDMS concentration and several geophysical parameters was unsuccessful (Kettle et al.,1999 ).

Quantitative studies that use data bases spanning roughly three decades (Simo and Dachs,2002; Simo and Vallina, 2007; Vallina et al., 2007 ) have concluded that the DMS andsolar radiation have a high positive correlation on seasonal time scales for most of theocean, favouring a negative feedback on climate. Another study (Larsen, 2005 ), based ona conceptual model, proposed a positive feedback. A positive feedback would also imply


an anti-correlation between Total Solar Irradiance (TSI) and cloud cover. Indeed, a globalanti-correlation between TSI and oceanic low cloud cover has been found (Kristjanssonet al., 2002; Lockwood, 2005 ). Furthermore, case studies do not allow conclusive resultsregarding the sign of the solar radiation and DMS correlation (Kniventon et al., 2003;Toole and Siegel, 2004; Toole et al., 2006 ).

A more recent study on decadal relation between north and south high latitude con-centrations of methane sulphonic acid (MSA), associated exclusively to DMS and TSI,found that at the time scales coincident with the 22 years magnetic solar cycle, the north-MSA and TSI follow each other, favoring a negative feedback on local climate; but atthe time scales of the 11 years solar cycle for north and south, the DMS-TSI presentsan anti-correlation that has increased since the 1940’s favoring a positive feedback onlocal climate (Mendoza and Velasco, 2009 ), this indicates that the relations between thevariables of interest may change with time and location.

The purpose of the present study is to examine the relationship between DMS and climaticand solar phenomena, through low cloud cover anomaly (LCC), sea surface temperature(SST) and the solar ultraviolet radiation A (UVA) in a selected location and at time-scalesshorter than the solar cycle.

2 Region of Study and Data

The data analysis was performed for the Southern Hemisphere between 40o−60o latitudefor the entire strip length. Considering the abundance of chlorophyll by regions, this areaincludes the biogeochemical provinces SSTC, FKLD, SANT, ANTA, NEWS, CHIL andTASM (Longhurst, 1995 ).

We are interested in this area because it is the least polluted in the world, the so-calledpristine zone, then solar effects on biota and climate should be more evident. Over90% of the area is ocean, the remaining corresponds to the southern parts of Chile,Argentina, Tasmania and New Zealand (South Island). The studied period is 1983-2008, containing almost 25 year of data. The DMS data set was obtained from the sitehttp://saga.pmel.noaa.gov/dms. The original data series in this web site contains the DMSmeasurements collected in the global oceans during 1972-2010 and are given as raw datasamples (Fig.1a). The sequence of data presents important data gaps in space and time,more details are given in Kettle et al. (1999 ). We also use the SST time series (Fig. 1b),obtained from Climatic Research Unit http://www.cru.uea.ac.uk/cru/data/temperature/hadsst2sh.txt. Another time series we use is the low cloud cover anomaly data (LCC) fromInternational Satellite Cloud Climatology Project (ISCCP) http://isccp.giss.nasa.gov.Two series of low cloud cover anomaly were obtained: Visible-Infrared (VI-IR)and In-frared (IR) (Fig. 1c and 1d). Finally, we work with solar ultra violet radiation A data(UVA), which comprises wavelengths from 320 to 400 nm, because the 95% of wave-lengths longer than 310 nm reach the surface (Lean et al., 1997 ) and has a large impact


on marine ecosystems (Hader et al., 2003; Toole et al., 2006; Hefu et al. 1997; Slezaket al., 2003; Kniventon et al., 2003; Hader et al., 2011 ). We use the UVA compositeseries from the Nimbus-7 (1978-1985), NOAA-9 (1985-1989), NOAA-11 (1989-1992) andSUSIM satellites between 1992 and 2008 (Fig. 1e)(DeLand et al., 2008 ). The atmo-spheric attenuated UVA series was calculated at sea surface level using the Santa BarbaraDisorted Atmospheric Radiative Transfer (SBDART) program; it was obtained from thesite http://www.icess.ucsb.edu/esrg/SBDART.html (Ricchiazzi et al., 1998 ). The averageattenuation was estimated at about 5% less than the top of the atmosphere series.

3 The Method

Some of the previous efforts on elucidating a plausible contribution of DMS on the Earth’sclimate have been mostly based on correlation analysis models. Such analysis suggestcertain relation between the time series, however, they are of global nature and do notprovide a precise information about when such a relation occurs. Moreover, the factthat two series have similar periodicities does not necessarily imply that one is the causeand another is the effect. Besides, even if the correlation coefficient is very low, thereis the possibility that an existing relation could be of non-linear nature, or that there isa strong phase shift between both phenomena. Here we apply three different non-linearanalysis to study the time series: the Wavelet Method, the Fractal Analysis and the VectorAutoregressive Analysis. They are briefly described below.

3.1 Wavelet Analysis

In order to analyze local variations of power within a single non-stationary time seriesat multiple periodicities we apply the wavelet method using the Morlet function (Tor-rence and Compo, 1998; Grinsted et al., 2004 ). Meaningful periodicities (confidence levelgreater than 95%) must be inside the cone of influence (COI), which is the region of thewavelet spectrum outside which the edge effects become important (Torrence and Compo,1998 ). An appropriate background spectrum is either the white noise (with a flat powerspectrum) or the red noise (increasing power with decreasing frequency). We calculatethe significance levels in the global wavelet spectra with a simple red noise model (Gilmanet al., 1963 ). We only take into account those periodicities above the red noise level.

Furthermore, we find the wavelet coherence spectra (Torrence and Compo, 1998; Grinstedet al., 2004 ). It is especially useful in highlighting the time and frequency intervals whenthe two phenomena have a strong interaction. If the coherence of two series is high, thearrows in the coherence spectra show the phase between the phenomena. Arrows at 0o

(horizontal right) indicate that both phenomena are in phase, and arrows at 180o (hor-izontal left) indicate that they are in antiphase. It is very important to point out thatthese two cases imply a linear relation between the considered phenomena. Arrows at 90o

and 270o (vertical up and down, respectively) indicate an out of phase situation, whichmeans that the two phenomena have a non-linear relation.


We also include the global spectra in the wavelet and coherence plots, which is the averagepower at each period over the whole time span considered within the COI (e.g. Velascoand Mendoza, 2008 ).The uncertainties of the peaks of both global wavelet and coherencespectra are obtained from the peak full-width at the half-maximum of the peak.

3.2 Fractal Analysis

The time series can be characterized by a non-integer dimension (fractal dimension) whentreated as random walks or self-affine profiles. Self-affine systems are often characterizedby the roughness, which is defined as the fluctuation of the height over a length scale. Forself-affine profiles, the roughness scales with the linear size of the surface by an exponentcalled the roughness or Hurst exponent (H ). Self-affine fractals are generally treated quan-titatively using spectral techniques (Turcotte, 1992 ). Also, his referred to as the index ofdependence, and is the relative tendency of a time series either to regress strongly to themean or to cluster in a direction (Kleinow, 2002 ). Most natural systems follow a biasedrandom walk or trend with statistical noise that can be approximated as Gaussian.

The strength of the trend and the noise level can be measured by analyzing the ratioR/S, where R indicates the range and S the standard deviation of the observed values,on a time scale in term of H, that is, when H is above 0.50 (Hurst, 1951 ). A value of1 < H < 1.50 results in afractal dimension(Ds) nearby to a line. A persistent time seriesresults in a smooth line, with fewer peaks than a random walk. Antipersistent series with0 < H < 0.50 have a higher Ds. The H exponent and Ds are related as Ds = 2−H(Peters, 1997, Voss, 1988 ).

3.3 Vector Autoregressive Analysis (VAR)

The VAR is a tool for multivariate time series. In the VAR all variables are consideredendogenous, as each is expressed as a linear function of its own lagged values and thelagged values of the other considered variables. An univariate auto regression is a single-equation, single-variable linear model in which the current value of a variable is explainedby its own lagged values. The VAR is an n-equation, n-variable linear model in whicheach variable is in turn explained by its own lagged values, plus current and past valuesof the remaining n-1 variables. This simple framework provides a systematic way to cap-ture rich dynamics in multiple time series with a coherent and credible approach to datadescription, forecasting, structural inference and analysis (Sims, 1980 ).

The VAR is also a powerful technique for generating reliable forecasts in the short term,it has certain limitations, such as omitting the possibility to consider non-linear relation-ships between variables and it does not take into account conditional heteroscedasticityproblems (that is, the observations sample error variances are different from each other)or structural change in the estimated parameters. The VAR finds the relations among


a set of variables and is appropriate for forecasting variables in systems of interrelatedtime series, where each variable helps to predict the other variables. The VAR is also fre-quently used in analyzing the dynamic impact of different types of disturbances in systemsof variables (Lutkepohl, 2005 ). The estimated VAR model uses the method of ordinaryleast squares. In the methodology of autoregressive models the modeling depends on thebehavior of a variable (Yt) according to its own past values. The general expression of theVAR model is given by:

Yt = A1Yt+1 + . . .+ ApYt−p +Bxt + εt (1)

Where Yt is a vector of endogenous variables. To estimate the VAR model you selecta number of variables that make up the model (Lutkepohl, 2005 ). A basic part of theVAR is the use of an impulse-response function that shows the response of current andfuture values of each of the variables, in other words, this function traces the responseof the endogenous variables in the model to an impulse (Lin, 2006 ), we apply here thisparticular aspect of the VAR.

4 Results and Discussion

The results obtained with the previous analysis will be presented and discussed in thissection.

4.1 The Main Frequencies

The results of the wavelet method are shown in Fig. 1, a summary of the main periodicitiesfound for all the time series is shown in the left column of Table 1. Fig. 1e for the UVAshows that the 11 years periodicity is the most significant, although it is partially outsidethe COI due to the short time interval. The LCCIR and LCCVI-IR coincide in two (0.9and 5 years) out of three periodicities. To corroborate the frequencies obtained with thewavelet analysis we apply fractal analysis. The results of this method are in Fig. 2, also,a summary appears in the right column of Table 1. With few exceptions, we notice thatmost of the wavelet periodicities, considering the uncertainties, coincide with the fractalperiodicities. Also, according to the H coefficient all the time series can be predicted. Theexception is the UVA; we consider that this is due to the shortness of the series comparedto the 11 years dominant periodicity.

4.2 The Wavelet Coherence Spectra

The coherence wavelet spectra in Fig. 3 present the coherence analysis between DMSvs SST, DMS vs LCCIR, DMS vs LCCVI-IR and DMS vs UVA respectively along 1992-2008. We choose this time interval because the DMS time series has the largest quantity ofdata. For each panel, the time series appears at the top, the wavelet coherence spectrumappears at the middle and the global wavelet coherence spectra is at the right. Table 2


summarizes the results. Fig. 3a, shows that the DMS and SST time series have the mostpersistent and prominent coherence ∼ 4 years and tend to be in phase. The DMS andLCCIR time series in Fig. 3b present the most prominent and persistent coherence ∼ 5years, and tend to be in anti-phase. The DMS and LCCVI-IR time series show persistentcoherences ∼ 3 and 5 years, they tend to be in phase and anti-phase respectively. TheDMS and UVA time series show persistent coherence at ∼ 3 years but it is not veryprominent, in fact the prominent coherences are between ∼ 0.4 and 1.2 years, they arevery localized in time and tend to be in anti-phase.

There is predominance in the periodicity between 3 and 5 years. Peaks shorter than1 year may be due to seasonal climatic phenomena. The ∼ 2 years period can be associ-ated with the Quasi-Biennial Oscillation (QBO) in the stratosphere(Holton et al., 1972;Dunkerton, 1997; Baldwin et al., 2001; Naujokat, 1986; Holton et al., 1980 )and withthe solar activity (Kane, 2005 ).The periodicities ∼ 3 and 4 years could be related tothe El Nino-Southern Oscillation (ENSO) (Nuzhdina, 2002; Njau, 2006: Enfield, 1992;Trenberth et al., 1997; Philander, 1990 ) that is a large-scale climatic phenomenon. Theperiodicities ∼ 5 years can be a harmonic of the 11 years solar cycle (Djurovic andPaquet, 1996 ). From Fig. 3 and Table 2, we notice a consistent correlation between DMSand SST and an anti-correlation between DMS and UVA, the relation between DMS andclouds is mainly non-linear. The anti-correlation between UVA and DMS suggest a posi-tive feedback, as discussed in other works (Larsen, 2005 ) or as implied by the findings ofother papers (Mendoza and Velasco, 2009; Lockwood, 2005; Kristjansson et al., 2002 ).

4.3 The VAR Impulse-Response Function

The impulse-response function appears in Fig. 4. In most cases we observe an impulse-response of 10 ± 2 months. In the present study the impulse-response function showsstatistically how the Earth’s climate has a recovery period of approximately 10 month.From top to bottom in Fig. 4 we notice the strongest meaningful responses of: the DMSto the impulse of SST, the LCC to DMS and the SST to the LCC. This indicates relationsamong clouds, DMS and SST and between the SST and DMS, supporting the existenceof relations among the elements of the feedback process hypothesized in Section 1.

5 Conclusions

Here we study relationship between DMS and the SST, LCCIR, LCCVI-IR and UVAusing different methods of analysis. The DMS, SST, LCCIR and LCCVI-IR series showpersistence and therefore have the possibility of a future estimation. For the UVA, theresults are not realistic and this is due to the shortness of the series that have prominentperiodicities for 11 years. The fractal analysis corroborated the wavelet periodicities.Using the wavelet method of spectral analysis, we found a predominance of periodicitybetween 3 and 5 years. The periodicities ∼ 3 and 4 years could be related to the ENSO.The periodicities ∼ 5 years are associated with solar activity. We found a correlation


between DMS and SST and an anticorrelation between DMS and UVA, the relation be-tween DMS and clouds is mainly non-linear. Then, our results suggest a positive feedbackinteraction among DMS, solar radiation and clouds. The VAR impulse-response functionmodel indicates statistically that Earth’s climate has a recovery period of approximately10 months and the existence of strong relations among low clouds, DMS and SST andbetween the SST and DMS.

6 References

Andreae, M.O., Crutzen, P.J. Atmospheric aerosols: biogeochemical sources and role inatmospheric chemistry, Science 276, 1052-1058, 1997.

Baldwin, M. P., Gray, L. J., Dunkerton, T. J., Hamilton, K., Haynes, P. H., Randel,W. J., Holton, J. R., Alexander, M. J., Hirota, I., Horinouchi, T., Jones, D. B., Kin-nersley, J. S., Marquardt, C., Sato, K., Takahashi, M., The Quasi-Biennial Oscillation,Reviews of Geophysics, Vol. 39, Num. 2, 179-229, 2001.

Charlson, R.J., Lovelock, J.E., Andreae, M.O., Warren, S.G. Oceanic phytoplankton,atmospheric sulfur, cloud albedo and climate: a geophysiological feedback. Nature 326,655-661, 1987.

Chen, T., Rossow, B. W., Zhang, Y., Radiative effects of Cloud-Type variations, Journalof climate Vol. 13, 1999.

DeLand, M. T., Cebula, R. P., Creation of a composite solar ultraviolet irradiance dataset. J. Geophys. Res 113, A11103, 2008.

Djurovic, D., Pquet, P., The common oscillations of solar activity, the geomagnetic field,and the earth’s rotation, Solar Physics, Vol. 167, 427-439, 1996.

Dunkerton, T. J., The role of gravity waves in the quasi-biennial oscillation. J. Geo-phys. Res., 102, 26,053-26,076, 1997.

Enfield, D. B., Historical and prehistorical overview of El Nio/Southern Oscillation. Cam-bridge, Cambridge University Press, 1, 95-117, 1992.

Grinsted, A., Moore, J., Jevrejera, S., Application of the cross wavelet transform andwavelet coherence to geophysical time series. Nonlinear Processes in Geophysics 11, 561-566, 2004.

Gunson, J.R., Spall, S.A., Anderson, T.R., Jones, A., Totterdell, I.J., Woodage, M.J., Cli-mate sensitivity to ocean dimethylsulphide emissions. Geophys. Res. Lett. 33, L07701,2006.


Hder, D. P., Helbling, E. W., Williamson, C. E., Worrest, R. C., Effects of UV radia-tion on aquatic ecosystems and interactions with climate change, Photochem. Photobiol.Sci. 10, 242-260, 2011.

Hder, D. P., Kumar, R. C., Smith, R. C., Worrest, R. C., Aquatic ecosystems: effects ofsolar ultraviolet radiation and interactions with other climatic change factors, Journal ofthe Royal Society of Chemistry, 2, 39-50, 2003.

Hefu, Y., Kirst, G. O., Effects of UV radiation on DMS content and DMS formationof Phaeocystis Antartica, Polar Biol. 18, 402-409, 1997.

Holton, J.R., Lindzen, R. S., An updated theory for the Quasi-Biennial cycle of thetropical stratosphere. Journal of the Atmospheric Sciences Vol. 29, 1076-1080. 1972.

Holton, J.R., Tan, H. C., The influence of the equatorial Quasi-Biennial Oscillation onthe global atmospheric circulation at 50mb, Journal of Atmospheric Science., Vol. 37,2200-2208, 1980.

Hurts, H.E., Long-term storage capacity of reservoirs, Trans. Am. Soc. Civ. Eng.,116, 770-779, 1951.

Kettle, A.J., Andreae, M. O., Amouroux, D., et al. A global data base of sea surfaceDimethylsulfide (DMS) measurements and a procedure to predict sea surface DMS as afunction of latitude, longitude and month, Global Biogeochemical Cycles 13, 399-444,1999.

Kleinow, T., Testing Continuous Time Models in Financial Markets, Berlin Humboldt-Univ., Diss. 1, 2002

Kniventon, D. R., Todd, M. C., Sciare, J., Mihalopoulos, N., Variability of atmosphericdimethylsulphide over the southern Indian Ocean due to changes in ultraviolet radiation,Global Biogeochem. Cycles 17, 1096, 2003.

Kristjnsson, J. E., Staple, A., Kristiansen, J., Kaas, E., A new look at possible con-nection between solar activity, clouds and climate, Geophys. Res. Lett., 29, 2107, 2002.

Larsen, S.H., Solar variability, dimethylsulphide, clouds and climate, Global Biogeochem.Cycles,19, GB1014, 2005.

Lean, J. R., Lee, G. J., Woods, H., Hickey, T. N., Puga, J., Detection and parameteri-zation of variations in solar mid- and near-ultraviolet radiation (200-400 nm). Geophys.Res. Lett. 102, 939-956, 1997.


Lin, J. L., Teaching notes on impulse response function and structural VAR, Instituteof Economics National Chengchi University, 2006.

Lockwood, M., Solar outputs, their variations and their effects on Earth, in The Sun,Solar Analogs and the Climate, Saas-Fee Advanced Course 34, 109-306, Springer, theNetherlands, 2005.

Longhurst, A., Sathyendranath, S., Platt, T., Caverhill, C., An estimate of global primaryproduction in the ocean from satellite radiometer data, J. Plankton Res., 17, 1245-1271,1995.

Ltkepohl, H., New introduction to multiple time series analysis, Springer-Verlag, 2005.

Mendoza, B., Velasco, V., High-Latitude Methane Sulphonic Acid Variability and So-lar Activity. J. Atm. and Solar-Terr Phys. 71, 33-40, 2009.

Miller, A. J., Alexander, M.A., et al. Potential feedbacks between Pacific Ocean ecosys-tems and interdecadal climate variations. Bull. Am. Meteorol. Soc., 84, 617-633, 2003.

Naujokat, B., An update of the observed Quasi-Biennial Oscillation of stratospheric windsover the tropics, Journal of the Atmospheric Sciences, Vol. 43, 1873-1877, 1986.

Njau, E. C., Solar activity, El Nio-Southern oscillation and rainfall in India, PakistanJournal of Meteorology, Vol. 3, 2006.

Nuzhdina, M. A., Connection between ENSO phenomena and solar and geomagneticactivity, Natural Hazards and Earth System Sciences, 2, 83-89, 2002.

Peters, E.E., Chaos and order in the capital markets. John Wiley and sons, 1997.

Philander, S. G., El Nio, La Nia and the Southern Oscillation, San Diego: AcademicPress, 1990.

Ricchiazzi, P., Yang, S., Gautier, C., Sowle, D., SBDART:A Research and teaching soft-ware tool for plane-parallel radiative transfer in the Earth’s atmosphere. Bull. Amer.Meteor. Soc. 79, 2101-2114, 1998.

Rossow, W. B., Schiffer, R. A., Advances in understanding clouds from ISCCP, Bull.Am. Meteor. Soc. 80, 2261- 2287, 1999.

Sarmiento, J. L., Gruber, N., Brzezinski, M. A., Dunne, J. P., High-latitude controlsof thermocline nutrients and low latitude biological productivity. Nature, 427 (6969),


56-60, 2004.

Shaw, G.E., Benner, R.L., Cantrell, W., Veazey, D., The regulation of climate: A sulfateparticle feedback loop involving deep convection-An editorial essay. Climate Change 39,23-33, 1998.

Simo, R., Production of atmospheric sulfur by oceanic plankton: biogeochemical, eco-logical and evolutionary links, Trends in Ecology and Evolution, 16, 287-294, 2001.

Simo, R., Dachs J., Global ocean emission of dymethylsulfide predicted from biogeo-physical data, Global Biogeochem. Cycles, 16, 2002.

Simo, R., Vallina, S. M., Strong relationship between DMS and the solar radiation doseover the global surface ocean, Science, 315, 506-508, 2007.

Sims, C.A., Macroeconomics and Reality, Econometrica 48, 1-48, 1980.

Slezak, D., Herndl, G. J., Effects of ultraviolet and visible radiation on the cellular con-centration of dimethylsulfoniopropionate (DMSP) in Emilianiahuxleyi (strain L), Mar.Ecol. Prog. Ser. 246, 61-71, 2003.

Solanki, S. K., Solar variability and climate change: is there a link?, Astronomy andGeophysics, Vol. 43,509-513, 2002.

Toole, D. A., Slezak, D., Kiene, R. P., Kieber, D. J., Effects of solar radiation on dimethyl-sulfide cycling in the western Atlantic Ocean, Deep-Sea Res. I, 53, 136-153, 2006.

Toole, D.A., Siegel D. A., Light-driven cycling of Dimethylsulfide (DMS) in the Sar-gasso Sea: Closing the Loop, Geophys. Res. Lett., Vol. 31, 2004.

Torrence, C., Compo, G. B., A practical guide to wavelet analysis, Amer. Meteor.Soc.79, 61-78, 1998.

Trenberth, K. E., The definition of El Nio, Bull. Am. Meteor. Soc. 78 (12): 2771-7, 1997.

Turcotte, D.L., Fractals and Chaos in Geology and Geophysics, Cambridge UniversityPress, 1992.

Vallina, S. M., Sim, R., Gass, S., DeBoyer-Montgut, C. D., Jurado, E., Dachs, J., Analysisof a potential ”solar radiation dose-dimethylsulfide-cloud condensation nuclei” link fromglobally mapped seasonal correlations, Global Biogeochemical Cycles 21, GB2004, 2007.


Velasco, V., Mendoza, B., 2008. Assessing the relationship between solar activity andsome large scale climatic phenomena, Advances in Space Research 42, 866-878, 2008.

Voss, R.J., Fractals in Nature: from characterization to simulation, The Science of fractalimages, Heinz-Otto Peitgen and Dietmar Saupe, Eds., New York: Springer-Verlag, 21-70,1988.

7 Figure Captions

Fig. 1 Wavelet analysis. For each panel at the top there is the time series, at the middlethe wavelet spectra and at the right the global wavelet spectra. (a) Dimethylsulfide (DMS).(b) Sea Surface Temperature (SST). (c) Low Cloud Cover Infrared Anomaly (LCCIR).(d) Low Cloud Cover Visible-Infrared Anomaly (LCCVI-IR). (e) Ultraviolet RadiationA (UVA). The gray code indicating the statistical significance level for the spectral plotsappears at the bottom of the figure; in particular the 95% level is inside the black contours.The dashed lines in the global wavelet spectra represent the red noise level.

Fig. 2 Power Spectrum plot for the time lapse 1983-2008, the linear trend has beenremoved. The power-spectral density is given as a function of frequency for time scales ofmonths. The fitted line is used to estimate the fractal dimension (Ds). (a) Dimethylsulfide(DMS).(b) Sea Surface Temperature (SST). (c) Low Cloud Cover Anomaly (LCCIR). (d)Low Cloud Cover Anomaly (LCCVI-IR). (e) Ultraviolet Radiation A (UVA).

Fig. 3 Wavelet coherence analysis. For each panel at the top there is the time series,at the middle the wavelet coherence spectra and at the right the global wavelet coherencespectra. (a) DMS (pointed line) and SST (dashed line). (b) DMS (pointed line) andLCCIR (dashed line). (c) DMS (pointed line) and LCCVI-IR. (d) DMS (pointed line)and UVA. The gray code indicating the statistical significance level for the spectral plotsappears at the bottom of the figure; in particular the 95% level is inside the black contours.

Fig. 4 Estimate of the VAR impulse-response function of the various series studiedhere for 10 month in the autoregressive model. The dashed lines are the uncertainties ofthe prediction.


8 Figures

Figure 1: Wavelet Analysis.


Figure 2: Power Spectrum Analysis.


Figure 3: Wavelet Coherence Analysis.


Figure 4: Impulse-Response Function.


Documents

Possible climate interaction mechanism between terrestrial climatic parameters and solar activity