Regional Air Temperature Research Articles

Decadal climate variability in the Mediterranean region: roles of large-scale forcings and regional processes

We analyze decadal climate variability in the Mediterranean region using observational datasets over the period 1850–2009 and a regional climate model simulation for the period 1960–2000, focusing in particular on the winter (DJF) and summer (JJA) seasons. Our results show that decadal variability associated with the winter and summer manifestations of the North Atlantic Oscillation (NAO and SNAO respectively) and the Atlantic Multidecadal Oscillation (AMO) significantly contribute to decadal climate anomalies over the Mediterranean region during these seasons. Over 30% of decadal variance in DJF and JJA precipitation in parts of the Mediterranean region can be explained by NAO and SNAO variability respectively. During JJA, the AMO explains over 30% of regional surface air temperature anomalies and Mediterranean Sea surface temperature anomalies, with significant influence also in the transition seasons. In DJF, only Mediterranean SST still significantly correlates with the AMO while regional surface air temperature does not. Also, there is no significant NAO influence on decadal Mediterranean surface air temperature anomalies during this season. A simulation with the PROTHEUS regional ocean–atmosphere coupled model is utilized to investigate processes determining regional decadal changes during the 1960–2000 period, specifically the wetter and cooler 1971–1985 conditions versus the drier and warmer 1986–2000 conditions. The simulation successfully captures the essence of observed decadal changes. Model set-up suggests that AMO variability is transmitted to the Mediterranean/European region and the Mediterranean Sea via atmospheric processes. Regional feedbacks involving cloud cover and soil moisture changes also appear to contribute to observed changes. If confirmed, the linkage between Mediterranean temperatures and the AMO may imply a certain degree of regional decadal climate predictability. The AMO and other decadal influences outlined here should be considered along with those from long-term increases in greenhouse gas forcings when making regional climate out-looks for the Mediterranean 10–20 years out.

Origins and Levels of Seasonal Forecast Skill for Sea Ice in Hudson Bay Using Canonical Correlation Analysis

Abstract Canonical correlation analysis (CCA) is used to estimate the levels and sources of seasonal forecast skill for July ice concentration in Hudson Bay over the 1971–2005 period. July is an important transition month in the seasonal cycle of sea ice in Hudson Bay because it is the month when the sea ice clears enough to allow the first passage of ships to the Port of Churchill. Sea surface temperature (quasi global, North Atlantic, and North Pacific), Northern Hemisphere 500-mb geopotential height (z500), sea level pressure (SLP), and regional surface air temperature (SAT) are tested as predictors at 3-, 6-, and 9-month lead times. The model with the highest skill has three predictors—fall North Atlantic SST, fall z500, and fall SAT—and significant tercile forecast skill covering 61% of the Hudson Bay region. The highest skill for a single-predictor model is from fall North Atlantic SST (6-month lead). Fall SST explains 69% of the variance in July ice concentration in Hudson Bay and a possible atmospheric link that accounts for the lagged relationship is presented. CCA diagnostics suggest that changes in the subpolar North Atlantic gyre and the Atlantic multidecadal oscillation (AMO), reflected in sea surface temperature, precedes a deepening/weakening of the winter upper-air ridge northwest of Hudson Bay. Changes in the height of the ridge are reflected in the strength of the winter northwesterly winds over Hudson Bay that have a direct impact on the winter ice thickness distribution; anomalies in winter ice severity are later reflected in the pattern and timing of spring breakup. July ice concentration in Hudson Bay has declined by approximately 20% per decade between 1979 and 2007, and the hypothesized link to the AMO may help explain this significant loss of ice.

Investigating the possible causes of climate change in India with satellite measurements

India's climate is changing gradually from one year to another. A significant increase in surface air temperature in different Indian regions was observed from 1998 onwards, which became all the more pronounced in 2009. Monthly mean maximum and minimum temperatures were higher than normal values over the whole of India, accompanied by a sharp increase in the duration and frequency of hot days and heatwave conditions. In the light of these observations the paper is aimed at investigating the spatial and temporal variation of air temperature at different altitudes and the interannual trend for different months over the Indian region for a period of 108 years from 1901 to 2008. It is observed that the variation in air temperature is not uniform at all places in India. The yearly mean air temperature for the period 1948–2008 exhibits an increasing trend at all atmospheric layers from the surface up to 10 km. The recently observed increase in surface air temperature and longer periods of heatwaves and hot days in India may be due to a higher occurrence of dust storms and forest fires and the influence of local factors, such as a sharp population growth accompanied by uncontrolled urbanization and rapid industrialization, resulting in high levels of pollution and imbalances in the regional climate.

Evaluation of snow cover and depth simulated by a land surface model using detailed regional snow observations from Austria

An evaluation is undertaken of the accuracy with which the Joint UK Land Environment Simulator (JULES) can simulate snow cover and depth when driven using data from the Hadley Centre regional climate model. The JULES model provides the facility to diagnose the thermal and hydrological state of the land surface and soil given time‐varying inputs of air temperature, wind speed, humidity, shortwave and longwave radiation, and precipitation. The observed data set used in this study consists of daily snow depths measurements at 601 climate stations with more than 15 years of observations in the period from January 1976 to December 2000. In this study, the JULES model was driven using two data sets at 25 km horizontal resolution: one produced using the U. K. Met Office Hadley Centre regional climate model HadRM3‐P (RCM), the other in which regional climate model precipitation and air temperature data were replaced with observed values (RCM+PT). The results indicate good agreement between the land surface model simulations and observations of snow cover at climate stations. The median snow cover accuracy indices for all 601 stations were 89% and 91% for the RCM and the combined RCM+PT driving data sets, respectively, with only a small interannual variation. In contrast, the differences between modeled and measured snow depth were much larger. The median values of mean snow depth bias were similar, −0.4 cm for RCM and −1.2 cm for RCM+PT; however, the RCM simulation was found to overestimate the observed snow depth at more than 25% of climate stations. The extent to which the results from RCM‐driven simulations match observed data is strongly related to the accuracy of the RCM precipitation. The large overestimation has significant impact on the snow mass simulation and the assessment of extreme values in the mountains. We note that even if snow cover can be simulated with a high degree of accuracy, this should not imply a similarly high degree of accuracy in the simulation of snow depth. Model performance was poorest in regions of significant topographic heterogeneity and our findings suggest that the most promising additional model developments should be directed toward computationally efficient representations of subgrid topography.

Empirical modeling of spatial and temporal variation in warm season nocturnal air temperatures in two North Idaho mountain ranges, USA

Accurate, fine spatial resolution predictions of surface air temperatures are critical for understanding many hydrologic and ecological processes. This study examines the spatial and temporal variability in nocturnal air temperatures across a mountainous region of Northern Idaho. Principal components analysis (PCA) was applied to a network of 70 Hobo temperature loggers systematically distributed across 2 mountain ranges. Four interpretable modes of variability were observed in average nighttime temperatures among Hobo sites: (1) regional/synoptic; (2) topoclimatic; (3) land surface feedback; (4) canopy cover and vegetation. PC time series captured temporal variability in nighttime temperatures and showed strong relationships with regional air temperatures, sky conditions and atmospheric pressure. PC2 captured the topographic variation among temperatures. A cold air drainage index was created by predicting PC2 loadings to elevation, slope position and dissection indices. Nightly temperature maps were produced by applying PC time series back to the PC2 loading surface, revealing complex temporal and spatial variation in nighttime temperatures. Further development of both physically and empirically based daily temperature models that account for synoptic atmospheric controls on fine-scale temperature variability in mountain ecosystems are needed to guide future monitoring efforts aimed at assessing the impact of climate change.

Asynchrony of spring phytoplankton response to temperature driver within a spatial heterogeneity bay of Three-Gorges Reservoir, China

The increased air temperature is expected to have important driver on spring phytoplankton dynamics. To test whether spatial heterogeneity modifies the synchronous responses of phytoplankton to regional temperature driver, we evaluate temporal coherences for physical factors (temperature, water stability and non-algal light extinction), nutrients (nitrogen, phosphorus and silicon), and biomass and density of phytoplankton by Pearson correlation analysis and synchrony for phytoplankton community dynamics by Mantel test and nonmetric multi-dimensional scaling (NMS), during spring bloom (February 23–April 28, 2005) within Xiangxi Bay, a high spatial gradient bay of Three-Gorges Reservoir (China). The high level of temporal coherences for surface water temperature ( r = 0.946, p < 0.01) and relative water column stability ( r = 0.750, p < 0.01) were found between pair sites (A and B), in which the increase trends occurred with increase in regional air temperature during the study period. However, the low synchrony for phytoplankton dynamics were indeed observed between Site A and B, especially for the density of common dominant taxa ( Cyclotella spp.: r = 0.155, p = 0.388) and community structure (Mantel test: r = 0.351). Moreover, the local habitat characteristics such as nutrient (nitrogen and phosphorus) and non-algal light extinction showed low levels of temporal coherence. It indicated that local community of phytoplankton varies rather independently within the single lentic bay with high spatial heterogeneity and that dispersal of algal organisms among locations cannot overwhelm out these local dynamics. Contrary to many studies, the present results argued that, in a small geographic area (i.e., a single reservoir bay of approximately 24 km length), spatial gradients also may influence spring phytoplankton response to regional temperature driver.

A Simple Method to Evolve Daily Ground Temperatures from Surface Air Temperatures in Snow-Dominated Regions

AbstractA simple model is developed to evolve daily ground temperatures from surface air temperatures (SATs) in snow-dominated areas. Ground surface temperatures (GSTs) are calculated by propagating the daily SAT through the snowpack, and attenuating the signal amplitude. Subsequent subsurface heat transfer is then modeled using the analytical solution of the one-dimensional heat conduction equation. The thermal impacts of nonconductive heat transfer processes and seasonal freeze thaw are implicitly represented by the time-dependent apparent thermal diffusivity of the subsurface. The model is tested in four snow-dominated regions: Barrow, Council, Ivotuk (all in Alaska) and Reynolds Creek Experimental Watershed (in Idaho). The model captures the seasonal evolution of the ground temperature at all sites. The model demonstrates the feasibility of simulating subsurface temperatures using only air temperature and snow depth.

Space observations of inland water bodies show rapid surface warming since 1985

Surface temperatures were extracted from nighttime thermal infrared imagery of 167 large inland water bodies distributed worldwide beginning in 1985 for the months July through September and January through March. Results indicate that the mean nighttime surface water temperature has been rapidly warming for the period 1985–2009 with an average rate of 0.045 ± 0.011°C yr−1 and rates as high as 0.10 ± 0.01°C yr−1. Worldwide the data show far greater warming in the mid‐ and high latitudes of the northern hemisphere than in low latitudes and the southern hemisphere. The analysis provides a new independent data source for assessing the impact of climate change throughout the world and indicates that water bodies in some regions warm faster than regional air temperature. The data have not been homogenized into a single unified inland water surface temperature dataset, instead the data from each satellite instrument have been treated separately and cross compared. Future work will focus on developing a single unified dataset which may improve uncertainties from any inter‐satellite biases.

Effect of permafrost degradation on hydrological processes in typical basins with various permafrost coverage in Western China

Monthly discharge of four rivers with various permafrost coverage and little anthropogenic influence was used to identify effects of permafrost degradation during the last 50 years, which has occurred because of significant increases in air temperature in the river regions. The basins of the Shule, Heihe, Shiyang and upper Yellow Rivers in northwestern China have 73%, 58%, 33% and 43% permafrost coverage, respectively. There is snow cover in the basins and no rain to supply rivers during winter. The monthly recession coefficient (RC) in winter reflects groundwater conditions. The RC has increased obviously for the Shule and Heihe rivers with 73% and 58% permafrost coverage, respectively, but did not increase for the Shiyang River, and decreased insignificantly for the upper Yellow River, which had less permafrost coverage. There is a distinct positive relationship between RC and annual negative degree-day temperature (NDDT) at the meteorological stations in the basins with high permafrost coverage. These results imply that permafrost degradation due to climate warming affects hydrological processes in winter. The effect is obvious in the basins with high permafrost coverage but negligible in those with low permafrost coverage. Permafrost degradation increases infiltration, enlarges the groundwater reservoir, and leads to slow discharge recession. The result means that hydrological processes are affected strongly by permafrost degradation in river basins with high permafrost coverage, but less in river basins with less permafrost coverage.

Non-climatic thermal adaptation: implications for species' responses to climate warming

There is considerable interest in understanding how ectothermic animals may physiologically and behaviourally buffer the effects of climate warming. Much less consideration is being given to how organisms might adapt to non-climatic heat sources in ways that could confound predictions for responses of species and communities to climate warming. Although adaptation to non-climatic heat sources (solar and geothermal) seems likely in some marine species, climate warming predictions for marine ectotherms are largely based on adaptation to climatically relevant heat sources (air or surface sea water temperature). Here, we show that non-climatic solar heating underlies thermal resistance adaptation in a rocky-eulittoral-fringe snail. Comparisons of the maximum temperatures of the air, the snail's body and the rock substratum with solar irradiance and physiological performance show that the highest body temperature is primarily controlled by solar heating and re-radiation, and that the snail's upper lethal temperature exceeds the highest climatically relevant regional air temperature by approximately 22°C. Non-climatic thermal adaptation probably features widely among marine and terrestrial ectotherms and because it could enable species to tolerate climatic rises in air temperature, it deserves more consideration in general and for inclusion into climate warming models.

Thermal state of permafrost in North America: a contribution to the international polar year

AbstractA snapshot of the thermal state of permafrost in northern North America during the International Polar Year (IPY) was developed using ground temperature data collected from 350 boreholes. More than half these were established during IPY to enhance the network in sparsely monitored regions. The measurement sites span a diverse range of ecoclimatic and geological conditions across the continent and are at various elevations within the Cordillera. The ground temperatures within the discontinuous permafrost zone are generally above −3°C, and range down to −15°C in the continuous zone. Ground temperature envelopes vary according to substrate, with shallow depths of zero annual amplitude for peat and mineral soils, and much greater depths for bedrock. New monitoring sites in the mountains of southern and central Yukon suggest that permafrost may be limited in extent. In concert with regional air temperatures, permafrost has generally been warming across North America for the past several decades, as indicated by measurements from the western Arctic since the 1970s and from parts of eastern Canada since the early 1990s. The rates of ground warming have been variable, but are generally greater north of the treeline. Latent heat effects in the southern discontinuous zone dominate the permafrost thermal regime close to 0°C and allow permafrost to persist under a warming climate. Consequently, the spatial diversity of permafrost thermal conditions is decreasing over time. Copyright © 2010 Crown in the right of Canada and John Wiley &amp; Sons, Ltd.

Observational Evidence of the Impact of Vegetation Cover on Surface Air Temperature Change in China

AbstractUsing NOAA/AVHRR normalized difference vegetation index (NDVI) data, the impact of vegetation cover on surface air temperature change is investigated by subtracting the reanalysis from the observation (observation minus reanalysis, OMR). The results show that the temperature OMR trend is inversely correlated to the surface vegetation index and the strong (weak) surface warming response to the surface barrenness (greenness). Areas with NDVI less than 0.1 show a large warming trend of more than 0.2°C/10a, while areas with NDVI large than 0.5 show little surface warming and even decreasing trends of temperature. Moreover, the OMR trend sensitively responds to the seasonal vegetation cover change. Atmosphere‐land feedbacks over areas with different vegetation covers can alter the response of regional surface air temperature to global warming induced by the increase in anthropogenic greenhouse gases concentrations. Thus, we suggest that the prediction of future change of climate in China should incorporate the impacts of spatial‐temporal variability in vegetation cover.

Sensitivity of arctic summer sea ice coverage to global warming forcing: towards reducing uncertainty in arctic climate change projections

Substantial uncertainties have emerged in Arctic climate change projections by the fourth Intergovernmental Panel on Climate Change assessment report climate models. In particular, the models as a group considerably underestimate the recent accelerating sea ice reduction. To better understand the uncertainties, we evaluated sensitivities of summer sea ice coverage to global warming forcing in models and observations. The result suggests that the uncertainties result from the large range of sensitivities involved in the computation of sea ice mass balance by the climate models, specifically with the changes in sea ice area (SIA) ranging from 0.09 × 10 6 to −1.23 × 10 6 km 2 in response to 1.0 K increase of air temperature. The sensitivities also vary largely across ensemble members in the same model, indicating impacts of initial condition on evolution of feedback strength with model integrations. Through observationally constraining, the selected model runs by the sensitivity analysis well captured the observed changes in SIA and surface air temperatures and greatly reduced their future projection uncertainties to a certain range from the currently announced one. The projected ice-free summer Arctic Ocean may occur as early as in the late 2030s using a criterion of 80% SIA loss and the Arctic regional mean surface air temperature will be likely increased by 8.5 ± 2.5 °C in winter and 3.7 ± 0.9 °C in summer by the end of this century.

Increased greenhouse gases enhance regional climate response to a Maunder Minimum

The climate responses to the Maunder Minimum‐type total solar irradiance (TSI) decrease in the pre‐industrial (PI) era and IPCC B1 global warming scenario are examined using the NCAR CAM3 coupled with a mixed‐layer slab ocean model. The TSI reduction shifts the AO/NAO state to a more negative index, resulting in regional surface air temperature (SAT) change in the model. In addition, the positive sea ice‐solar radiation feedback amplifies the surface cooling due to the TSI decrease. The global annual average cooling effect induced by the TSI decrease is reduced to 0.254°C in the B1 scenario from 0.347°C in the PI epoch, this difference being due to both the suppressed sea ice‐solar radiation feedback and the stronger greenhouse effect associated with the increase of greenhouse gases. However, regional SAT changes in the B1 scenario are strengthened markedly due to the enhanced negative AO/NAO. The warming in the western Greenland, Central Asia and East Asia will be enhanced in the B1 global warming scenario when the TSI decreases. These differences in response between the B1 and PI scenarios illustrate that the impact of multi‐decadal solar variation on climate depends on the atmospheric background trace gas composition. A future solar grand minimum does not offset the CO2‐induced warming in the B1 scenario, but through the enhanced negative AO/NAO may indeed amplify the warming in certain regions.

Sensitivity of arctic summer sea ice coverage to global warming forcing: towards reducing uncertainty in arctic climate change projections

Substantial uncertainties have emerged in Arctic climate change projections by the fourth Intergovernmental Panel on Climate Change assessment report climate models. In particular, the models as a group considerably underestimate the recent accelerating sea ice reduction. To better understand the uncertainties, we evaluated sensitivities of summer sea ice coverage to global warming forcing in models and observations. The result suggests that the uncertainties result from the large range of sensitivities involved in the computation of sea ice mass balance by the climate models, specifically with the changes in sea ice area (SIA) ranging from 0.09 × 106 to −1.23 × 106 km2 in response to 1.0 K increase of air temperature. The sensitivities also vary largely across ensemble members in the same model, indicating impacts of initial condition on evolution of feedback strength with model integrations. Through observationally constraining, the selected model runs by the sensitivity analysis well captured the observed changes in SIA and surface air temperatures and greatly reduced their future projection uncertainties to a certain range from the currently announced one. The projected ice-free summer Arctic Ocean may occur as early as in the late 2030s using a criterion of 80% SIA loss and the Arctic regional mean surface air temperature will be likely increased by 8.5 ± 2.5 ◦C in winter and 3.7 ± 0.9 ◦C in summer by the end of this century.

Hydrological and geomorphological significance of rock glaciers in the dry Andes, Chile (27°–33°S)

AbstractThe latitudinal‐altitudinal distribution of rock glaciers in the dry Chilean Andes between 27° and 33°S was analysed and their hydrological and geomorphological significance evaluated. Statistical estimation techniques were used based on digital elevation models and air photo interpretation, and sediment budget models were applied to assess surface‐lowering rates. The estimated specific density of rock glaciers in the study area is 1.4 per cent, which corresponds to 147.5 km2 and a water equivalent of 2.37 km3. A comparison with glacier water equivalents derived from revised glacier inventory data shows that rock glaciers are more significant stores of frozen water than glaciers between 29° and 32°S. The latitudinal‐altitudinal distribution of rock glacier limits roughly follows the mean annual air 0°C isotherm, but in the southern part of the study area it extends into elevations where modern regional mean annual air temperatures exceed 0°C. High Andean surface‐lowering rates inferred from rock glacier sediment budget models are in the order of 0.6–0.7 mm yr−1, which is comparable to previous results from the dry Chilean Andes, but lower than in the more humid Andes further south. Copyright © 2009 John Wiley &amp; Sons, Ltd.

Intraseasonal non-stationarity of the leading modes of atmospheric moisture over Europe during summer

A gridded monthly precipitable water (PW) data for 1979–2007 from the NCEP/NCAR reanalysis are used to investigate summertime interannual PW variability over Europe and its relation to the key climate parameters in the region. During summer season the first EOF mode of PW, explaining 27–41% of its total variance, demonstrates significant month-to-month changes in its structure, thus, implying its essential non-stationarity. The second EOF mode of PW is also non-stationary during the summer season. In contrast to precipitation, both leading modes of PW are not associated with the North Atlantic Oscillation (NAO), as well as with other regional teleconnections, suggesting relatively minor role of the atmospheric dynamics in atmospheric moisture variability over Europe during summer season. Analysis of links between leading EOF modes of regional PW and air temperature (AT) has revealed a strong link between PW and AT over Europe, persisting during entire summer season. Locally, these links imply that positive (negative) AT anomalies result in enhanced (decreased) PW over particular region. Revealed links between leading modes of PW and AT highlight important role of thermodynamics in summertime PW variability over Europe. Detected relatively weak and unstable links between leading modes of PW and precipitation over Europe were somewhat expected since in contrast to atmospheric moisture, regional precipitation variability is largely driven by the atmospheric dynamics (particularly, the NAO).

Air temperature and thermal comfort conditions in mountainous and urban regions

The present study explored air temperature and thermal comfort conditions in mountainous region of Nafpaktia (Municipality of Apodotia, Prefecture of Aitoloakarnania), a region without important commercial, industrial or other activities in west continental Greece, and in Athens, a metropolitan urban center in southeast continental Greece. There were 11 study sites in the case of Nafpaktia (676–1455 m altitude), areas with different altitude, ground cover, plant species and plant density. There was one study site in Athens (30 m altitude) nearby a densely built area, under the pressure of industrial activities and heavy traffic. Air temperature and humidity, 1.5 m above ground surface, were monitored simultaneously every 15 min in each site for the period between 1 July and 31 August 2006. Air temperature and relative humidity data were used for the calculation of the thermohygrometric index (THI) from which thermal comfort conditions were evaluated. For this work, two sites of mountainous Nafpaktia region (MNR), one characterized by the lower (fir and chestnut forest area) and the other by the higher (Evinos River area) average THI value, are presented. Results showed that MNR provided clearly improved air temperature and thermal comfort conditions in comparison to Athens in the study period. Fir and chestnut forest area was more beneficial, from a bioclimatological point of view, compared to Evinos River area. Therefore, fir and chestnut forest area should be further supported as a reliable alternative tourist destination for rest and recreation.

Spatiotemporal change in China’s climatic growing season: 1955–2000

The timing, length, and thermal intensity of the climatic growing season in China show statistically significant changes over the period of 1955 to 2000. Nationally, the average start of the growing season has shifted 4.6-5.5 days earlier while the average end has moved 1.8-3.7 days later, increasing the length of the growing season by 6.9-8.7 days depending on the base temperature chosen. The thermal intensity of the growing season has increased by 74.9-196.8 growing degree-days, depending on the base temperature selected. The spatial characteristics of the change in the timing and length of the growing season differ from the geographical pattern of change in temperatures over this period; but the spatial characteristics of change in growing degree-days does resemble the pattern for temperatures, with higher rates in northern regions. Nationally, two distinct regimes are evident over time: an initial period where growing season indicators fluctuate near a base period average, and a second period of rapidly increasing growing season length and thermal intensity. Growing degree-days are highly correlated with March-to-November mean air temperatures in all climatic regions of China; the length of the growing season is likewise highly correlated with March-to-November mean air temperatures except in east, southeast and southwest China at base temperature of 0A degrees C and southeast China at base temperature of 5A degrees C. The growing season start date appears to have the greater influence on the length of the growing season. In China, warmer growing seasons are also likely to be longer growing seasons.

Urban heat island diagnosis using ASTER satellite images and ‘in situ’ air temperature

This study demonstrates that thermal satellite images combined with ‘in situ’ ground data can be used to examine models of heat island genesis and thus identify the main causes of urban heat islands (UHIs). The models, although proposed over 30 years ago, have not been thoroughly evaluated due to a combination of inadequate ground data and the low resolution of thermal satellite data. Also there has been limited understanding of the relevance of satellite-derived surface temperatures to local and regional scale air temperatures. A cloud-free ASTER thermal image of urban and rural areas of Hong Kong was obtained on a winter night with a well-developed heat island, accompanied by a 148 km vehicle traverse of air temperatures. Over the whole traverse a high R 2 of 0.80 was observed between surface and air temperatures, with the two datasets showing a similar amplitude and general trend, but with the surface exhibiting much higher local variability than air temperature. Gradients in both surface and air temperature could be related to differences in land cover, with little evidence of large scale advection, thus supporting the population/physical structure model of UHI causation, rather than the advection model. However, the much higher surface and air temperatures observed over the largest urban area, Kowloon, than over any smaller urban centre with similar physical structure in the New Territories, would seem more indicative of the advection model. The image and ground data suggest that Kowloon's urban canopy layer climate is mainly influenced by local city structure, but it is also modified by a strongly developed, regional scale urban boundary layer which has developed over the largest urban centre of Kowloon, and reinforces heating from both above and below.