6.1 GLOBAL SOLAR RADIATION in a SOUTHERN AFRICAN SAVANNA ENVIRONMENT M. Nasitwitwi 1,*, W. G. Bailey 2, and L. J. B. Mcarthur 3
Total Page：16
File Type：pdf, Size：1020Kb
6.1 GLOBAL SOLAR RADIATION IN A SOUTHERN AFRICAN SAVANNA ENVIRONMENT M. Nasitwitwi 1,*, W. G. Bailey 2, and L. J. B. McArthur 3 1. Department of Geography, Douglas College, New Westminster, British Columbia, V3L 5B2 2. Department of Geography, Simon Fraser University, Burnaby, British Columbia, Canada, V5A 1S6 3. Experimental Studies Division, Meteorological Service of Canada, Downsview, Ontario, Canada, M3H 5T4 1.0 Introduction * described dry season effects of aerosols (Eck et al., A pervasive dearth of radiation knowledge exists 2001; Holben et al., 2001; Schafer et al. 2002) as a in Africa, with Zambia being a typical case. Amongst result of anthropogenic biomass burning over the rural the radiation budget components, global solar landscape (Desanker et al., 1997; Scholes et al., radiation has in past years assumed economic 1996). importance as a renewable energy alternative (Lewis, Angstrom-type empirical radiation models based 1981; World Solar Programme, 1999; Omran, 2000). on sunshine hours were earlier utilized to provide More recently, global solar radiation is being studied radiation data in Zambia without validation against due to its importance in providing energy for the actual measurements (Spain, 1971). Subsequent earth’s climate system. work refined this model using Zambian data National weather networks throughout Africa, (Mwangala and Mukambulo, 1980; Jain, 1983; Jain have not routinely observed radiation fluxes because and Jain, 1988). Further, Jain and Jain (1988) of the high cost of instruments and poor funding for proposed regional coefficients of the Angstrom-type the Meteorological Departments. In Zambia, this radiation model applicable to all Zambian locations. situation has been compounded by a severe attrition Errors in all these modelling procedures were due of qualified personnel to other countries; apathy largely to the low accuracy of radiation and sunshine amongst the scientific community towards instrumentation. In addition, Jain and Jain (1988) atmospheric research; and an absence of local isolated substantial unsystematic scatter of regression training facilities by which to develop a radiation cadre parameters of the Angstrom model indicative of (United Nations Educational Scientific and Cultural physical atmospheric processes that were beyond the Organization, 2001). Consequently, only a short resolution of Angstrom’s model. monthly archive of global solar radiation measured at Being tropical, the Zambian savanna ecosystem several Zambian stations between 1977 and 1983 experiences three seasons: a warm-wet season from forms the basis of Zambian literature published by December to March; a cool-dry season from April to Mwangala and Mukambulo (1980), Jain (1983) and July; and a hot-dry season from August to November Jain and Jain (1988). Fragmentary measurements (Archer, 1971). This ecosystem, locally known as were continued for a number of years, but have not miombo, is characterized by annual bushfires from been published because their reliability was strongly August to October, which generate substantial questioned (Lewis, 1981). Recent measurements of amounts of aerosols seasonally (Desanker et al., global solar radiation in western Zambia have 1997; Scholes et al., 1996). This paper discusses the properties of global * Corresponding author address: M. Nasitwitwi, solar radiation at Lusaka, Zambia measured over a Department of Geography, Douglas College, New wide range of meteorological conditions. Houghton’s Westminster, BC Canada V3L 5B2; e-mail address: [email protected] radiation model, which has a tradition of studying urban aerosols and cloud effects in mid-latitude total cloudless-sky transmission is the product of environments, is applied to a savanna landscape transmissions due to water vapour absorption Ψwa, characterized by tropical atmospheric dynamics and aerosol absorption Ψda, water vapour scattering Ψws, pyrogenic aerosols produced by anthropogenic Rayleigh scattering Ψrs, and aerosol scattering Ψds biomass burning. The model is used to determine the (Houghton, 1954; Davies et al., 1975). Houghton’s radiative influence of aerosols and other atmospheric (1954) parameterization assumes that absorption of controls on cloudless-sky global solar radiation the solar beam occurs before scattering; half of dust regimes in savanna environments. depletion is due to absorption; and that absorption due to ozone negligible (Davies et al., 1975). Direct- 2.0 Theory beam solar radiation transmitted after absorption and The sun emits electromagnetic energy in the scattering is therefore defined as wavelength range 0.1 to 4 µm. Extraterrestrial solar radiation (Kex) incident at the top of the atmosphere is So = Io cos Z Eo ψ wa ψ da ψ ws ψ rs ψ ds (3) determined by solar output, sun-earth positioning, latitude, time of the year and time of the day such that Under cloudless conditions, the amount of diffuse solar radiation Do reaching the earth’s surface is K I Z E (1) ex = o cos o defined as where Io is the solar constant, Z is the solar zenith (1 −ψ ws ψ rs ψ ds ) Do = Io cos Z Eo ψ wa ψ da (4) angle and Eo is the eccentricity correction factor of the 2 earth’s orbit. Between the top of the atmosphere and the Therefore, global solar radiation under cloudless sky earth’s surface, the atmosphere reflects, absorbs and conditions K↓o is the sum of the direct-beam solar scatters the incoming solar beam. The flux at the radiation So and diffuse solar radiation Do earth’s surface, termed global solar radiation, is the sum of the direct-beam and the diffuse components. K ↓o = So + Do (5) Assuming that multiple reflections between the (ψ ws ψ rs ψ ds + 1) K ↓o = Io cos Z Eo ψ wa ψ da . (6) surface and the atmosphere are non-existent, global 2 solar radiation K↓ is Parameterizations of various transmissions were K ↓ = t Kex (2) computed as follows: m where t is the clearness index. ψ da =ψ ds = 0.975 (7) When the direct beam from the sun penetrates m ψ d =ψ da ψ ds = K (8) the cloudless atmosphere, its intensity is reduced by ψ ws = 1 − 0.0225wm (9) scattering in which there is an angular redistribution of 0.3 the energy and by absorption in which the energy is ψ wa = 1 − 0.077wm (10) 2 3 4 converted into either heat or photochemical change. ψ rs =0.972 − 0.08262m + 0.00933m − 0.00095m + 0.0000437m Probability of non-interference of the incoming beam (11) by the atmosphere is called transmittance. In the broadband spectrum from 0.1 to 4 µm wavelengths, where the dust factor k = 0.95, w is precipitable water Hourly visibility observations were converted to (in cm) and m is optical air mass (Houghton, 1954; atmospheric turbidity following the procedure of Davies et al., 1975). MacClatchey and Selby (1972) 3.0 Experimental Procedure ≈ 3.91 ’ β = 0.55α∆ − 0.01162[0.02472(Vis − 5)+1.132] (13) From April to December 2000, K↓ was « Vis continuously measured at Lusaka, Zambia (15° 24′ S, 28° 19′ E, 1154m) using an Eppley PSP Precision where β is atmospheric turbidity, Vis is visibility (in pyranometer. The instrument was mounted on a kilometres) and α is the wavelength exponent equal to levelled 1.5 m platform on a rooftop at University of 1.3 (Iqbal, 1983). Zambia and was calibrated at 2.74 µV/Wm-2 by the Cloudless hours with Z ≤ 70° were analyzed by National Atmospheric Radiation Centre of Canada. A months to determine seasonal variations. Thus the Campbell Scientific 21X datalogger measured signals Houghton model estimated hourly K↓o in each month at one-second intervals and stored one-minute with hourly transmissions due to water vapour averages. Constant daily zero offsets given by absorption and scattering derived from daily values of averages of one-minute means of K↓ recorded precipitable water. Rayleigh scattering was calculated between zenith angles 102o (nautical twilight) and as a function of pressure-corrected air mass (Iqbal, 108o (astronomical twilight) were used to correct the 1983). A constant dust factor k = 0.95 was used to measurements. pre-evaluate transmission due to dust scattering and One-minute Kex values were computed from daily absorption. The residual between this model and values of the sun declination and the equation of time. measured hourly K↓o was subsequently used to The solar constant of 1366.1 Wm-2 in accordance with derive non-constant dust factors for the period April to the ASTM 460 standard of the U.S. Naval November 2000. Effects of precipitable water and dry Observatory and an eccentricity correction factor from molecular air were jointly determined by the shortfall McCullough’s (1968) were also employed. The one- between Kex and the hourly cloudless model at k = 1. minute Kex values were subsequently summed into In turn, aerosol influences were obtained by hourly and daily values. subtracting the hourly cloudless model with a non- Supplemental data obtained from the Zambian constant dust factor from the hourly cloudless model Meteorological Department at Lusaka City Airport at k = 1. Differencing the hourly cloudless model with included total cloud amount, dry and wet bulb a non-constant dust factor from actual surface temperatures, upper air soundings and horizontal measurements provided the seasonal trend of cloud visibility. Precipitable water was evaluated from effects. midday radiosonde upper air soundings of pressure, temperature and relative humidity. A model of 4.0 Results precipitable water w and near-surface dewpoint 4.1 Seasonal radiation variability -2 -1 temperature Td was used to interpolate for periods Kex ranged between 26 MJm d in the cool-dry that lacked soundings as follows: season and 41 MJm-2d-1 in the warm-wet season (Figure 1) as a consequence of seasonally varying ln w = 0.087 Td − 0.4963 . (12) solar zenith angles. Daily K↓ values were moderate in the cool-dry season with greater variability at the season’s onset.