Search Results
You are looking at 1 - 10 of 27 items for
- Author or Editor: J. R. Garratt x
- Refine by Access: All Content x
Abstract
The Zilitenkevitch relation for nocturnal boundary-layer (NBL) depth h in terms of scales u */f and L is necessarily a poor predictor of h when single-point values of surface fluxes are used. This is because the latter are poor estimates of ensemble-mean fluxes which satisfy this relation. The argument is equally valid both in an evolving and equilibrium NBL.
Abstract
The Zilitenkevitch relation for nocturnal boundary-layer (NBL) depth h in terms of scales u */f and L is necessarily a poor predictor of h when single-point values of surface fluxes are used. This is because the latter are poor estimates of ensemble-mean fluxes which satisfy this relation. The argument is equally valid both in an evolving and equilibrium NBL.
Abstract
An evaluation of the clear-sky longwave irradiance at the earth’s surface (LI) simulated in climate models and in satellite-based global datasets is presented. Algorithm-based estimates of LI, derived from global observations of column water vapor and surface (or screen air) temperature, serve as proxy “observations.” All datasets capture the broad zonal variation and seasonal behavior in LI, mainly because the behavior in column water vapor and temperature is reproduced well. Over oceans, the dependence of annual and monthly mean irradiance upon sea surface temperature (SST) closely resembles the observed behavior of column water with SST. In particular, the observed hemispheric difference in the summer minus winter column water dependence on SST is found in all models, though with varying seasonal amplitudes. The analogous behavior in the summer minus winter LI is seen in all datasets. Over land, all models have a more highly scattered dependence of LI upon surface temperature compared with the situation over the oceans. This is related to a much weaker dependence of model column water on the screen-air temperature at both monthly and annual timescales, as observed. The ability of climate models to simulate realistic LI fields depends as much on the quality of model water vapor and temperature fields as on the quality of the longwave radiation codes. In a comparison of models with observations, root-mean-square gridpoint differences in mean monthly column water and temperature are 4–6 mm (5–8 mm) and 0.5–2 K (3–4 K), respectively, over large regions of ocean (land), consistent with the intermodel differences in LI of 5–13 W m−2 (15–28 W m−2).
Abstract
An evaluation of the clear-sky longwave irradiance at the earth’s surface (LI) simulated in climate models and in satellite-based global datasets is presented. Algorithm-based estimates of LI, derived from global observations of column water vapor and surface (or screen air) temperature, serve as proxy “observations.” All datasets capture the broad zonal variation and seasonal behavior in LI, mainly because the behavior in column water vapor and temperature is reproduced well. Over oceans, the dependence of annual and monthly mean irradiance upon sea surface temperature (SST) closely resembles the observed behavior of column water with SST. In particular, the observed hemispheric difference in the summer minus winter column water dependence on SST is found in all models, though with varying seasonal amplitudes. The analogous behavior in the summer minus winter LI is seen in all datasets. Over land, all models have a more highly scattered dependence of LI upon surface temperature compared with the situation over the oceans. This is related to a much weaker dependence of model column water on the screen-air temperature at both monthly and annual timescales, as observed. The ability of climate models to simulate realistic LI fields depends as much on the quality of model water vapor and temperature fields as on the quality of the longwave radiation codes. In a comparison of models with observations, root-mean-square gridpoint differences in mean monthly column water and temperature are 4–6 mm (5–8 mm) and 0.5–2 K (3–4 K), respectively, over large regions of ocean (land), consistent with the intermodel differences in LI of 5–13 W m−2 (15–28 W m−2).
Abstract
There are numerous reports in the literature of observations of land surface temperatures. Some of these, almost all made in situ, reveal maximum values in the 50°–70°C range, with a few, made in desert regions, near 80°C. Consideration of a simplified form of the surface energy balance equation, utilizing likely upper values of absorbed shortwave flux (≈1000 W m−2) and screen air temperature (≈55°C), that surface temperatures in the vicinity of 90°–100°C may occur for dry, darkish soils of low thermal conductivity (≈0.1–0.2 W m−1 K−1). Numerical simulations confirm this and suggest that temperature gradients in the first few centimeters of soil may reach 0.5°–1°C mm−1 under these extreme conditions. The study bears upon the intrinsic interest of identifying extreme maximum temperatures and yields interesting information regarding the comfort zone of animals (including man).
Abstract
There are numerous reports in the literature of observations of land surface temperatures. Some of these, almost all made in situ, reveal maximum values in the 50°–70°C range, with a few, made in desert regions, near 80°C. Consideration of a simplified form of the surface energy balance equation, utilizing likely upper values of absorbed shortwave flux (≈1000 W m−2) and screen air temperature (≈55°C), that surface temperatures in the vicinity of 90°–100°C may occur for dry, darkish soils of low thermal conductivity (≈0.1–0.2 W m−1 K−1). Numerical simulations confirm this and suggest that temperature gradients in the first few centimeters of soil may reach 0.5°–1°C mm−1 under these extreme conditions. The study bears upon the intrinsic interest of identifying extreme maximum temperatures and yields interesting information regarding the comfort zone of animals (including man).
Abstract
There is direct evidence that excess net radiation calculated in general circulation models at continental surfaces [of about 11–17 W m−2 (20%–27%) on an annual basis is not only due to overestimates in annual incoming shortwave fluxes [of 9–18 W m−2 (6%–9%)], but also to underestimates in outgoing longwave fluxes. The bias in the outgoing longwave flux is deduced from a comparison of screen-air temperature observations, available as a global climatology of mean monthly values, and model-calculated surface and screen-air temperatures. An underestimate in the screen temperature computed in general circulation models over continents, of about 3 K on an annual basis, implies an underestimate in the outgoing longwave flux, averaged in six models under study, of 11–15 W m−2 (3%–4%). For a set of 22 inland stations studied previously, the residual bias on an annual basis (the residual is the net radiation minus incoming shortwave plus outgoing longwave) varies between 18 and −23 W m−2 for the models considered. Additional biases in one or both of the reflected shortwave and incoming longwave components cannot be ruled out.
Abstract
There is direct evidence that excess net radiation calculated in general circulation models at continental surfaces [of about 11–17 W m−2 (20%–27%) on an annual basis is not only due to overestimates in annual incoming shortwave fluxes [of 9–18 W m−2 (6%–9%)], but also to underestimates in outgoing longwave fluxes. The bias in the outgoing longwave flux is deduced from a comparison of screen-air temperature observations, available as a global climatology of mean monthly values, and model-calculated surface and screen-air temperatures. An underestimate in the screen temperature computed in general circulation models over continents, of about 3 K on an annual basis, implies an underestimate in the outgoing longwave flux, averaged in six models under study, of 11–15 W m−2 (3%–4%). For a set of 22 inland stations studied previously, the residual bias on an annual basis (the residual is the net radiation minus incoming shortwave plus outgoing longwave) varies between 18 and −23 W m−2 for the models considered. Additional biases in one or both of the reflected shortwave and incoming longwave components cannot be ruled out.
Abstract
Evidence is presented that the excess surface net radiation calculated in general circulation models at continental surfaces is mostly due to excess incoming shortwave fluxes. Based on long-term observations from 22 worldwide inland stations and results from four general circulation models the overestimate in models of 20% (11 W m−2) in net radiation on an annual basis compares with 6% (9 W m−2) for shortwave fluxes for the same 22 locations, or 9% (18 W m−2) for a larger set of 93 stations (71 having shortwave fluxes only). For annual fluxes, these differences appear to be significant.
Abstract
Evidence is presented that the excess surface net radiation calculated in general circulation models at continental surfaces is mostly due to excess incoming shortwave fluxes. Based on long-term observations from 22 worldwide inland stations and results from four general circulation models the overestimate in models of 20% (11 W m−2) in net radiation on an annual basis compares with 6% (9 W m−2) for shortwave fluxes for the same 22 locations, or 9% (18 W m−2) for a larger set of 93 stations (71 having shortwave fluxes only). For annual fluxes, these differences appear to be significant.
Abstract
Aspects of the land-surface and boundary-layer treatments in some 20 or so atmospheric general circulation models (GCMS) are summarized. In only a small fraction of these have significant sensitivity studies been carried out and published. Predominantly, the sensitivity studies focus upon the parameterization of land-surface processes and specification of land-surface properties—the most important of these include albedo, roughness length, soil moisture status, and vegetation density. The impacts of surface albedo and soil moisture upon the climate simulated in GCMs with bare-soil land surfaces are well known. Continental evaporation and precipitation tend to decrease with increased albedo and decreased soil moisture availability. For example, results from numerous studies give an average decrease in continental precipitation of 1 mm day−1 in response to an average albedo increase of 0.13. Few conclusive studies have been carried out on the impact of a gross roughness-length change—the primary study included an important statistical assessment of the impact upon the mean July climate around the globe of a decreased continental roughness (by three orders of magnitude). For example, such a decrease reduced the precipitation over Amazonia by 1 to 2 mm day−1.
The inclusion of a canopy scheme in a GCM ensures the combined impacts of roughness (canopies tend to be rougher than bare soil), albedo (canopies tend to be less reflective than bare soil), and soil-moisture availability (canopies prevent the near-surface soil region from drying out and can access the deep soil moisture) upon the simulated climate. The most revealing studies to date involve the regional impact of Amazonian deforestation. The results of four such studies show that replacing tropical forest with a degraded pasture results in decreased evaporation (≈ 1 mm day−1) and precipitation (1–2 mm day−1), and increased near-surface air temperatures (≈2 K).
Sensitivity studies as a whole suggest the need for a realistic surface representation in general circulation models of the atmosphere. It is not yet clear how detailed this representation needs to be, but even allowing for the importance of surface processes, the parameterization of boundary-layer and convective clouds probably represents a greater challenge to improved climate simulations. This is illustrated in the case of surface net radiation for Aniazonia, which is not well simulated and tends to be overestimated, leading to evaporation rates that are too large. Underestimates in cloudiness, cloud albedo, and clear-sky shortwave absorption, rather than in surface albedo, appear to be the main culprits.
There are three major tasks that confront the researcher so far as the development and validation of atmospheric boundary-layer (ABL) and surface schemes in GCMs are concerned:
(i) There is a need to as” critically the impact of “improved” parameterization schemes on WM simulations, taking into account the problem of natural variability and hence the statistical significance of the induced changes.
(ii) There is a need to compare GCM simulations of surface and ABL behavior (particularly regarding the diurnal cycle of surface fluxes, air temperature, and ABL depth) with observations over a range of surface types (vegetation, desert, ocean). In this context, area-average values of surface fluxes will be required to calibrate directly the ABL/land-surface scheme in the GCM.
(iii) There is a need for intercomparisons of ABL and land-surface schemes used in GCMS, both for one- dimensional stand-alone models and for GCMs that incorporate the respective schemes.
Abstract
Aspects of the land-surface and boundary-layer treatments in some 20 or so atmospheric general circulation models (GCMS) are summarized. In only a small fraction of these have significant sensitivity studies been carried out and published. Predominantly, the sensitivity studies focus upon the parameterization of land-surface processes and specification of land-surface properties—the most important of these include albedo, roughness length, soil moisture status, and vegetation density. The impacts of surface albedo and soil moisture upon the climate simulated in GCMs with bare-soil land surfaces are well known. Continental evaporation and precipitation tend to decrease with increased albedo and decreased soil moisture availability. For example, results from numerous studies give an average decrease in continental precipitation of 1 mm day−1 in response to an average albedo increase of 0.13. Few conclusive studies have been carried out on the impact of a gross roughness-length change—the primary study included an important statistical assessment of the impact upon the mean July climate around the globe of a decreased continental roughness (by three orders of magnitude). For example, such a decrease reduced the precipitation over Amazonia by 1 to 2 mm day−1.
The inclusion of a canopy scheme in a GCM ensures the combined impacts of roughness (canopies tend to be rougher than bare soil), albedo (canopies tend to be less reflective than bare soil), and soil-moisture availability (canopies prevent the near-surface soil region from drying out and can access the deep soil moisture) upon the simulated climate. The most revealing studies to date involve the regional impact of Amazonian deforestation. The results of four such studies show that replacing tropical forest with a degraded pasture results in decreased evaporation (≈ 1 mm day−1) and precipitation (1–2 mm day−1), and increased near-surface air temperatures (≈2 K).
Sensitivity studies as a whole suggest the need for a realistic surface representation in general circulation models of the atmosphere. It is not yet clear how detailed this representation needs to be, but even allowing for the importance of surface processes, the parameterization of boundary-layer and convective clouds probably represents a greater challenge to improved climate simulations. This is illustrated in the case of surface net radiation for Aniazonia, which is not well simulated and tends to be overestimated, leading to evaporation rates that are too large. Underestimates in cloudiness, cloud albedo, and clear-sky shortwave absorption, rather than in surface albedo, appear to be the main culprits.
There are three major tasks that confront the researcher so far as the development and validation of atmospheric boundary-layer (ABL) and surface schemes in GCMs are concerned:
(i) There is a need to as” critically the impact of “improved” parameterization schemes on WM simulations, taking into account the problem of natural variability and hence the statistical significance of the induced changes.
(ii) There is a need to compare GCM simulations of surface and ABL behavior (particularly regarding the diurnal cycle of surface fluxes, air temperature, and ABL depth) with observations over a range of surface types (vegetation, desert, ocean). In this context, area-average values of surface fluxes will be required to calibrate directly the ABL/land-surface scheme in the GCM.
(iii) There is a need for intercomparisons of ABL and land-surface schemes used in GCMS, both for one- dimensional stand-alone models and for GCMs that incorporate the respective schemes.
Abstract
The Cold Fronts Research Programme (CFRP) concentrated upon cold frontal systems (synoptic scale fronts) occurring in southeast Australia during late spring and early summer (November and December); many of these tend to the complex, with change lines associated with sea breezes, prefrontal squall lines and related cold outflows (Garratt et al.). The present paper extends the earlier work to include additional observations of frontal events from field phase 3 of the CFRP held in late 1984. In addition we distinguish between this frontal system (Type 1) and a frontal system (Type 2) occurring throughout summer, but predominantly late in the season, which is often dry with a single major change line identified as a mesoscale (coastal) front. 0bservations from the CFRP and other sources are described covering a wide range of frontal intensities, with frontal speeds varying between 5 and 25 m s−1.
The nature of the low level flow behind, and in the vicinity of, the active surface cold front (SCF) is described. In the coastal region both systems show significant falls in θ e across the SCF in the lowest 500 m or so, but only Type 2 has a corrresponding fall in θ. Type 1 shows gradual falls in θ throughout the prefrontal zone related to subcloud evaporative cooling. Wind observations for both types reveal a feeder flow of cold air towards the front, below 500–1000 m height and within about 50–100 km of the SCF. Together the data suggest that the surface cold front has the local structure of an unsteady gravity current.
Nevertheless the Type 1 SCF is identified with a synoptic scale (Southern Ocean) cold front having an associated large-scale, deep tropospheric three-dimensional flow configuration and so must be predominantly under large-scale control. This is consistent with the lack of any significant diurnal influence on the eastward movement of the SCF. In contrast such a diurnal influence is found for the Type 2 SCF (which form ahead of a synoptic scale cold front often only distinguishable as a cloud band on satellite imagery) with the implication of substantial mesoscale forcing.
Both the Type 2 SCF and prefrontal squall lines of the Type 1 system tend to traverse the coastal region between late morning and early evening, with squall-line activity tending to be concentrated near the coast. This implies a strong influence of the diurnal cycle of boundary-layer heating over the land, as suggested in the case of the Type 1 leading change line by Berson et al.
Abstract
The Cold Fronts Research Programme (CFRP) concentrated upon cold frontal systems (synoptic scale fronts) occurring in southeast Australia during late spring and early summer (November and December); many of these tend to the complex, with change lines associated with sea breezes, prefrontal squall lines and related cold outflows (Garratt et al.). The present paper extends the earlier work to include additional observations of frontal events from field phase 3 of the CFRP held in late 1984. In addition we distinguish between this frontal system (Type 1) and a frontal system (Type 2) occurring throughout summer, but predominantly late in the season, which is often dry with a single major change line identified as a mesoscale (coastal) front. 0bservations from the CFRP and other sources are described covering a wide range of frontal intensities, with frontal speeds varying between 5 and 25 m s−1.
The nature of the low level flow behind, and in the vicinity of, the active surface cold front (SCF) is described. In the coastal region both systems show significant falls in θ e across the SCF in the lowest 500 m or so, but only Type 2 has a corrresponding fall in θ. Type 1 shows gradual falls in θ throughout the prefrontal zone related to subcloud evaporative cooling. Wind observations for both types reveal a feeder flow of cold air towards the front, below 500–1000 m height and within about 50–100 km of the SCF. Together the data suggest that the surface cold front has the local structure of an unsteady gravity current.
Nevertheless the Type 1 SCF is identified with a synoptic scale (Southern Ocean) cold front having an associated large-scale, deep tropospheric three-dimensional flow configuration and so must be predominantly under large-scale control. This is consistent with the lack of any significant diurnal influence on the eastward movement of the SCF. In contrast such a diurnal influence is found for the Type 2 SCF (which form ahead of a synoptic scale cold front often only distinguishable as a cloud band on satellite imagery) with the implication of substantial mesoscale forcing.
Both the Type 2 SCF and prefrontal squall lines of the Type 1 system tend to traverse the coastal region between late morning and early evening, with squall-line activity tending to be concentrated near the coast. This implies a strong influence of the diurnal cycle of boundary-layer heating over the land, as suggested in the case of the Type 1 leading change line by Berson et al.
Abstract
Results of recent turbulence sensor comparison experiments suggest that much of the source of data scatter in CDN (V) plots and of the systematic differences between data sets is due to calibration uncertainties associated with sensor performance in the field. The effects (if any) of fetch, wind duration and unsteadiness remain obscured in this experimental data scatter.
Vertical transfer of momentum over land may be described in terms of an effective roughness length or geostrophic drag coefficient which incorporates the effects of both friction and form drag introduced by flow perturbation around uneven topographical features.
Typically low relief topography and low mountains (peaks <0.5–1 km) require a geostrophic drag coefficient CDN ≈ 3×10−3, while land surfaces in general require CDN ≈ 2×10−3 for which CDN (10)≈ 10×10−3 and the effective aerodynamic roughness length ẑ 0(eff)≈ 0.2 m. The latter values satisfy, very approximately, the requirement of global angular momentum balance.
Abstract
Results of recent turbulence sensor comparison experiments suggest that much of the source of data scatter in CDN (V) plots and of the systematic differences between data sets is due to calibration uncertainties associated with sensor performance in the field. The effects (if any) of fetch, wind duration and unsteadiness remain obscured in this experimental data scatter.
Vertical transfer of momentum over land may be described in terms of an effective roughness length or geostrophic drag coefficient which incorporates the effects of both friction and form drag introduced by flow perturbation around uneven topographical features.
Typically low relief topography and low mountains (peaks <0.5–1 km) require a geostrophic drag coefficient CDN ≈ 3×10−3, while land surfaces in general require CDN ≈ 2×10−3 for which CDN (10)≈ 10×10−3 and the effective aerodynamic roughness length ẑ 0(eff)≈ 0.2 m. The latter values satisfy, very approximately, the requirement of global angular momentum balance.
Abstract
No abstract available.
Abstract
No abstract available.
Abstract
At an International Turbulence Comparison Experiment (ITCE) in Australia (1976), wind, temperature and humidity profiles, plus vertical fluxes of momentum, sensible heat and latent heat were measured for a limited range of unstable conditions, but with a wide variety of instrumentation.
Comparisons of like instruments, and of flux and profile measurements via the Monin-Obukhov similarity theory, have been used to assess systematic effects. The scatter about conventional flux-profile formulations varies with choice of stability parameter (z/L or Ri) and this also proves an effective means of identifying the source of the scatter. The cup anemometer results exhibit evidence of errors in calibration factors which are the major source of scatter in the unsmoothed measured gradients.
The limited stability range prevents unambiguous solution for all constants in the conventional flux-profile relationships. For momentum transfer, use of Ri as a stability parameter gives least sensitivity to error; adoption of a Φ M (Ri) of Pruitt et al. (1973) leads to k M = 0.33 ± 0.03, significantly below their value of 0.42. Adoption of the Φ M (z/L) of Businger et al. (1971) or Dyer (1974) leads to kM = 0.38 ± 0.04 (cf. their values of 0.35 and 0.41, respectively).
For sensible heat transfer values of kH are significantly lower than those of other workers, while for latent heat kW values similar to previous workers are obtained-this implies a kH < kW for ITCE 1976.
Abstract
At an International Turbulence Comparison Experiment (ITCE) in Australia (1976), wind, temperature and humidity profiles, plus vertical fluxes of momentum, sensible heat and latent heat were measured for a limited range of unstable conditions, but with a wide variety of instrumentation.
Comparisons of like instruments, and of flux and profile measurements via the Monin-Obukhov similarity theory, have been used to assess systematic effects. The scatter about conventional flux-profile formulations varies with choice of stability parameter (z/L or Ri) and this also proves an effective means of identifying the source of the scatter. The cup anemometer results exhibit evidence of errors in calibration factors which are the major source of scatter in the unsmoothed measured gradients.
The limited stability range prevents unambiguous solution for all constants in the conventional flux-profile relationships. For momentum transfer, use of Ri as a stability parameter gives least sensitivity to error; adoption of a Φ M (Ri) of Pruitt et al. (1973) leads to k M = 0.33 ± 0.03, significantly below their value of 0.42. Adoption of the Φ M (z/L) of Businger et al. (1971) or Dyer (1974) leads to kM = 0.38 ± 0.04 (cf. their values of 0.35 and 0.41, respectively).
For sensible heat transfer values of kH are significantly lower than those of other workers, while for latent heat kW values similar to previous workers are obtained-this implies a kH < kW for ITCE 1976.
