Info

aTemperature change for the five-century long period (K). bTemperature change during the twentieth century (K).

aTemperature change for the five-century long period (K). bTemperature change during the twentieth century (K).

work by Huang et al. (2000) in order to optimally detect the Northern Hemisphere century-long climatic trends for the past five centuries. In the works by Pollack et al. (1998) and Huang et al. (2000) the hemispheric and global GST trends were estimated by a simple arithmetic averaging of the individual GST reconstructions. According to Mann et al. (2003), the estimation of the whole Northern Hemisphere trend by simple averaging of the sparse, irregularly distributed borehole data, when most of them are situated in the extratropical belt, may lead to the noticeable bias. The use of the "optimal detection" technique (see Section 3.4.4) can increase the significance of the estimated climate change. For easier comparison the results of the individual GST reconstructions serving as input data were averaged on the 5° X 5° grid, which is generally used for the representation of SAT data (Jones et al., 1999). Authors have shown, however, that in their case the procedure of the optimal detection was insensitive to whether the 453 individual GST reconstructions or 94 grid boxes were used for the calculations. The data existed for the 94 grid boxes with 1-22 boreholes per box. The relative error weighting scheme based on occupancy of given grid box was employed. The borehole GST reconstructions can be represented as (M X N) matrix B, where M = 94 (number of grid boxes) and N = 6 (number of estimated time values: 1500,1600,1700,1800,1900, and 1980). Because the average logging date is 1978, the recent GST trends in this work are representatives of the 1900-1980 period and not of the whole twentieth century.

Applying the singular value decomposition (SVD) of the un-normalized, time-centered data matrix (see Section 2.3.4) one can express the matrix B in an empirical orthogonal eigenvector basis:

where X2t is the relative variance resolved by the ith eigenvector, and u. and v. are its normalized spatial and temporal patterns, respectively. According to Mann et al.'s (2003) estimates, only two eigenvectors appeared to be statistically significant. Thus, e.g., the optimal expression for the spatial pattern of the GST trend in a given century can be represented as a linear function of two spatially centered normalized eigenvectors u1 and u2:

where i is the spatially centered, normalized pattern in the given century, a and b are regression coefficients, and e is the residual error term. The estimation of the coefficients a and b in this case represents simple linear regression procedure (for the details of calculus see Mann et al., 2003).

Figure 101 shows comparison of the five-century GST trends for the Northern Hemisphere calculated by Huang et al. (2000) with the areally weighted and averaged borehole reconstruction by Mann et al. (2003) and Rutherford and Mann (2004) (www.ncdc.noaa.gov/pub/data/paleo/borehole/mann2003). The shift of the latter curve occurs because of the difference in the reference periods. While the Huang et al.'s (2000) reconstruction is referenced to the present, Mann et al. (2003) used 1900-1980 projection interval (the mean of 1900-1980 was set to zero) in their calculations. Both results are

Fig. 101. Comparison of the century-averaged climatic trends for the Northern Hemisphere (Huang et al., 2000) and spatially weighted and averaged borehole reconstruction by Mann et al. (2003).

generally coherent and exhibit five-century warming trend. Estimated warming is somewhat reduced in the amplitude for the latter reconstruction (0.97 K against 1.02 K). As shown in Figure 101, sixteenth to eighteenth century trends practically coincide for both reconstructions. The recent warming in the Huang et al.'s (2000) case is more regularly distributed over the last two centuries, while in the Mann et al.'s (2003) case a bit greater part of it occurred in the twentieth century (0.5 K warming for the whole century and 0.6 K during only 1900-1980 in the former and in the latter reconstructions, respectively). The reason for such discrepancy is that Huang et al.'s (2000) reconstruction does not include significant warming of the last two decades of the twentieth century. This end-of-century bias is the characteristic feature of many GST reconstructions using borehole T-z profiles logged before 1980. Similar coincidence was revealed between reconstructed GST and SAT trends as well as with the proxy-based estimates of the hemispheric climatic trends in the past centuries (Mann et al., 2003; see also the next section).

The most recent analysis by Pollack and Smerdon (2004) was primarily concerned with the problems of data aggregation, their gridding, and occupancy-dependent weighting. Averaging of the individual borehole data on the regular grid helps the local noise suppression (see Section 2.4.5). Authors have used 695 individual borehole GST reconstructions and have shown that the averaged five-century Northern Hemisphere and/or global mean trends did not depend essentially on the applied scheme of weighting and aggregation. Independent of the choice of the signal detecting technique obtained in this work large-scale means are generally consistent and equal to approximately 1K for the wide range of gridding and weighting schemes used. Authors also investigated spatial correlation between GST and SAT for different grid sizes. As shown, the correlation appears already at the 5° X 5° gridding base. However, most of the grid boxes of 5° size contained three or even less boreholes. This quantity is insufficient for an effective smoothing of the site-specific noise in the box-averaged GST histories. Thus, the low-occupation elements can weaken and/or obscure the spatial GST-SAT correlation. Authors have shown that the correlation can be increased by excluding the low-occupation grid elements from the dataset. Remaining high-occupancy grid boxes (>3 boreholes) yield a statistically significant GST-SAT correlation. Trends obtained for the 5°- and 30°-size datasets are practically similar. The overweighting influence of the grid boxes with the data clustering is of less importance.

A little bit different global climatic history of the past five centuries was obtained by Beltrami (2002), who performed GST reconstructions for 826 borehole temperature logs using SVD inversion technique. Later, similar work by Beltrami and Bourlon (2004) was based on the measurements performed at 558 boreholes distributed between 30° and 60°N of the Northern Hemisphere. As in the all above-cited works, T-z profiles for these investigations were taken from the Global Database of Borehole Temperatures. Their spatial distribution and global coverage are generally similar to the previous works. The best coverage is characteristic for North American continent, and approximately 50% of the temperature logs originate from Canada. As in the work by Huang et al. (2000), all temperature-depth profiles were inverted individually and inferred GST histories were then averaged using area-weighted technique. This procedure is indispensable in order to avoid exaggerated representation of the areas with noticeable abundance of boreholes (e.g., some regions of Canada). The latest temperature log was measured in 1999, and the authors have extrapolated calculated global climatic trend to the year 2000. Figure 102 shows globally averaged GST change for the last five centuries and similar result for the Northern Hemisphere.

Fig. 102. Mean global GST history derived from SVD inversion of 826 temperature logs (data by Beltrami, 2002) and mean GST history for the Northern Hemisphere between 30 and 60°N (latitudinal average based on 5° X 5° grid boxes; Beltrami and Bourlon, 2004).

2 - Northern Hemisphere between 30-60°N

1500 1600 1700 1800 1900 2000 YEAR, A.D.

The course of the GST history is generally coherent with that detected in the early works except of the provisional cooling in the 1900-1950 (1880-1930 in the later work by Beltrami and Bourlon, 2004) interval that was not found in other works. Total GST warming for the past 500 years, which is equal to ~0.9K, thus coincides well with the value determined by Huang et al. (2000). Independent estimate by Beltrami (2002) has shown that the global average ground heat flow has also increased together with average global GST. According to the study, the total heat absorbed by the ground during the recent 50 years amounts to 7.1 X 1021J; thus, it is comparable with the quantities absorbed by the whole atmosphere and/or by the continental glaciers. Cooling event in the end of the nineteenth and first half of the twentieth century has manifested itself as a decrease of ground temperature by 0.2-0.4K. It was not synchronous over the globe. It is likely the reason that it appears weaker and with phase shift on the global mean in comparison with the Northern Hemisphere GST average. In the Northern Hemisphere this cooling was detected over both North America and Central Europe. This cold event is well visible on the meteorological records from Prague, Vienna, and Munich and on the diagram of occurrences of GST extremes in the territory of the Czech Republic (Figure 75 of the previous section). The reason that this cold excursion was not recovered by other reconstructions (e.g., by Huang et al., 2000) lies probably in the use of simple arithmetic averaging in the above works. If this finding will be corroborated by further studies, it can inspire to reconsider the course of the last 500-year warming trend. According to the results by Beltrami and Bourlon (2004), half of the detected ~1K past five centuries' warming occurred in the recent 50 years.

Somewhat different magnitude of temperature changes derived from the GST reconstructions in comparison with climatic trends defined from proxy records has inspired recent discussions and attempts to re-assess the skill of the "borehole" method to draw out past GST changes from T-z profiles. The most recent attempt was undertaken in the work by González-Rouco et al. (2006) and has proved the power of borehole inversion technique to reconstruct long-term climatic trends and/or robustness of existing borehole database to reproduce past SAT variations (for details see Section 2.4.4, Chapter 2).

The investigations of the several century regional cumulative GST trends in Russia including Siberia are of special interest. The reasons are:

1. Their important role in the climate system both on the hemispheric and the global scales. According to Cohen et al. (2001), the Siberian high-pressure anomaly is a dominant forcing factor of the winter climate variability in the Northern Hemisphere.

2. Because extensive continental regions in Siberia are relatively weakly affected by the postindustrial anthropogenic activity, the GST trends reconstructed for this region can significantly contribute to ongoing debates over the human influence on the climate (see also Section 3.4).

The recent work by Pollack et al. (2003) is concentrated on the past five-century climate changes in the territory of Russia as derived from borehole thermometry. One hundred and one T-z profiles were selected for the GST reconstructions. The investigated sites are generally concentrated in the Ural Mts., NE and SE Siberia. The most significant amount (66 temperature logs) was measured from the southern frontier to polar region of the Ural Mts. Four sites are in the European part of Russia. Figure 103 shows examples of

Fig. 103. Average GST history of the past five centuries for the Ural Mts., SW and NE Siberia, and "all-Russia" (drawn from the data by Pollack et al., 2003). For comparison similar trend is also shown for North America (30-65°N, 65-125°W). The twentieth century GST trends are compared with those indicated by the SAT observations.

Fig. 103. Average GST history of the past five centuries for the Ural Mts., SW and NE Siberia, and "all-Russia" (drawn from the data by Pollack et al., 2003). For comparison similar trend is also shown for North America (30-65°N, 65-125°W). The twentieth century GST trends are compared with those indicated by the SAT observations.

cumulative GST trends for investigated territories as well as averaged for "all-Russia" (Pollack et al., 2003). For comparison, similar data for North America (30-65°N, 65-125°W) and the twentieth century Northern Hemisphere SAT trends are also shown. Results for the Ural Mts. are more significant for the entire Russia. They were calculated from the most robust dataset of 66 boreholes (65% of the Russian data) and show the best coincidence with the GST trend detected for North America. As for the North American reconstructions, the course of the five-century climate is general warming accelerated during the twentieth century. The warming in the recent 100 years reached 0.68 K. This value exceeds the cumulative warming in the four previous centuries. Strong warming trend of 0.22K/100 years is also characteristic for the nineteenth century. Thus, approximately 80% of the total five-century warming took place in the last 200 years, and over 60% of this warming occurred in the last century. As shown in Figure 103, such general climatic course is also characteristic for the "all-Russia" average GST history. Warming trends obtained for the 500-year period coincide well with the quantities calculated for North America, indicating similar climatic course over both extensive northern continents. The rate of the last 100-year warming coincides in North America and the Urals.

The data measured in SW Siberia are relatively sparse and come from different terrains. However, the century-long GST trends calculated from the SW Siberian boreholes are similar to the quantities obtained for the Urals. The total warming is somewhat weaker and is distributed more regularly over the past five centuries. The recent warming is also weaker. However, it is comparable in its magnitude with the warming rate indicated for the Urals region. Here it represents 55% of the total five-century climate warming. An extensive territory of the NE Siberia is represented by the data measured in only 13 holes. All investigated boreholes are drilled in the permafrost. The mean annual surface temperature in these sites remains below 0°C, and the temperature log exhibits below zero values throughout the whole interval of the measurements. The dependence of the permafrost in bedrock on the past and presently ongoing climate variations as well as the possibility of the GST history reconstruction from the temperature logs measured in such environment have been discussed in Section 2.8. Because of the relatively small number of borehole results averaged over vast area, the range of uncertainty of the GST reconstructions for NE Siberia is fairly large; thus, the conclusions about the past climate changes are only of tentative character (Pollack et al., 2003). The course of the past five-century climate change in this area is rather distinct from other Russian territory. In contrast with other investigated regions, NE Siberia was exposed to the modest cooling during the sixteenth to eighteenth centuries. Warming with an inexpressive rate of 0.06K/100 years began here only in the nineteenth century. During the twentieth century this warming rate has increased by almost an order of magnitude to 0.67K/100 years. This value is comparable with the warming rates characteristic for both "all-Russia" ensemble and North America.

Figure 103 also shows a comparison of the twentieth century warming rates calculated from the GST data and the averaged Northern Hemisphere SAT observations (data by Jones et al., 1999; www.cru.uea.ac.uk). These data represent the 5° X 5° gridded temperature anomalies from the base period 1961-1990. The rates of the twentieth century warming for separate regions were calculated by Pollack et al. (2003) from the data by Jones et al. (1999). They equal to 1.32, 1.54, 0.88, and 1.33K/100 years for the Urals, SW Siberia, NE Siberia, and "all-Russia" territories, respectively. The distribution of meteorological stations in the area is sparse enough; thus, calculated SAT warming trends can be regarded only as the first-order estimate. For all regions the twentieth century SAT trends exceed the "geothermal" estimate. The reasons for this discrepancy may be the fact that most of boreholes were logged prior to 1983 and therefore do not archive the later climate changes and/or the well-known rapid attenuation of the temperature fluctuations in the first 50 m depth interval, so that temperature logs that begin at greater depth do not contain recent GST excursions.

Although Russian GST reconstructions prior to about 1500 A.D. exhibit expanded uncertainties, some important conclusions are still possible. The above results support the finding that the last century warming that occurred in the high latitudes of the Northern Hemisphere comprised Russia also. Here it is not confined to the Arctic and the northernmost regions only, but embraced significant territories up to 45-50°N latitudes. Generally 70-80% of the observed warming occurred in the last 100 years. While the past climatic trends inferred for North America generally suggest one-way story and the twentieth century warming represents simple continuation of the previous long-term warming trend, the Russian GST reconstructions show the twentieth century warming as a continuation of a trend that began only at the start of the nineteenth century.

All cited works collectively illustrate that the reconstruction of the century-long trends derived from borehole data over long time periods represents independent and robust source of the paleoclimatic information and may be a useful complementary source to the SAT estimates as well as to traditional proxy reconstructions. Independent estimates of hemispheric and global ground temperature trends over the past five centuries from the information archived in borehole temperature profiles confirm the conclusion of the previous section that the late twentieth century warming is anomalous in a long-term context. Temporal resolution of these GST estimates decreases back in time. Thus, the meaningful comparison of the obtained long-term GST trends with high-resolution temperature estimates based on proxy climate data seems to be very useful.

3.3 Correlation Between GST Climate Reconstruction, Meteorological Data, and Proxies

Paleoclimate reconstructions describing surface temperature course on the hemispheric and/or global scale for the timescales over one to two millennia are generally obtained from traditional proxies as well as from the "geothermal" GST reconstructions. The conformity between GST histories, SAT, and proxy time series was already discussed in Sections 2.5 and 2.6 (Chapter 2). It will be captured in more detail in the present section. It should be declared from the very beginning that the conformity does not mean the identity and/or equality of the investigated sources. It is clear that they are not identical. The question is rather: do these sources show similar trends over long periods of time?

The instrumental record alone is clearly not enough to represent few centuries or longer climatic trends. Proxy and borehole paleoclimate reconstructions fill this vacuum. Borehole data are usually interpreted as reflecting true annual temperatures, incorporating all seasons at least on the longer timescales. They may be disturbed by various environmental factors, e.g., change in the land-use practices. However, as shown by Harris and Chapman (2001), this noise can be suppressed by a simple averaging of reduced T-z profiles. The major problems occurring for all proxy indicators are their chronological dating, calibration, and detection of what they really measure. For example, tree-ring records of climate change are an important component of multiproxy records used to infer climate variations, and all Northern Hemisphere temperature reconstructions use at least some tree-ring data. They provide powerful tools for the detection of past temperature variations; they are widespread, well dated, have high resolution, and extend the estimates of climatic variability well beyond the instrumental period. These series are generally calibrated assuming that they reflect annual variations. However, tree growth is more influenced by summer than winter temperatures. Further uncertainties can be introduced by the possible adaptation of the living species to new climate conditions and by expectable nonlinear response to the varying climate on the long timescales (for details see Section 1.2.3). Even if the paleoclimatologists are aware of the shortcomings, the clear meaning of the proxy record is not always described in many research works. To construct credible pattern of the past climate variations the scientists need to compare each other's interpretations and assumptions. The compilation of a heap of climatic histories seems to be more promising than the use of the single reconstructions.

Obviously, the integrated climate history does not represent a simple superposition of different proxy reconstructions and needs a complex technique that consolidates information that different sources contain. An overview of available large-scale climate reconstruction techniques and/or computer codes can be found, e.g., on the web site of the Climate Change Research Section (CCR) of the National Center for Atmospheric Research (www.cgd.ucar.edu/ccr/amman/millennium/recon). This and similar web pages describe the existing multiproxy climate reconstruction methods and provide the corresponding codes that can help to reproduce the individually published reconstructions as well as to assess the behavior of the given method and evaluate its strength and shortcomings. More detailed examination of the calculus for the multiproxy compilation especially for the noisy data can be found in the works by von Storch et al. (2004) and McIntyre and McKitrick (2005a), as well as in the subsequent discussions (Huybers, 2005; McIntyre and McKitrick, 2005b, c; von Storch and Zorita, 2005).

Below we describe some of the most noticeable examples from the growing amount of the past centuries' climate reconstructions that merge high-resolution proxy and/or instrumental data with the GST histories. Combination of only two different paleocli-matic data sources appears to be a simpler procedure than the multi-data reconstructions. Two-source merging can be performed by regression of the one source to another. For example, numerous reconstructions were performed by coupling of the instrumental SAT series with the lower resolution proxy indicators (Jones et al., 1998; Mann et al., 1998, 1999). Regression models are generally constructed for the more recent time interval where SAT and proxy data overlap and then extend to remaining proxy sections. Results of such extrapolation, however, imply that the reply of the proxy indicator to the changing climate remains invariable during the long period of time and the coupling of the proxy and SAT over the long-time interval is the same as it was in the period of the instrumental observations. Such simplification holds the source for possible bias.

Merging of two equally long climatic series appears to be more promising. Such kind of joint two-source past climate reconstruction was performed by Beltrami and Taylor (1995), who combined borehole CST histories with oxygen isotope data from an ice core. Both measurement sites were closely located in the Canadian Arctic. Because of different resolution of the records (annual versus decadal and longer), such coupling appears to be quite useful and can help to increase the resolution of the GST reconstruction. While the borehole GST histories reflect true annual temperatures at least on the longer scale, the oxygen isotope ratios measured in the ice cores represent mainly summer conditions and summarize responses to various processes. Together with average air temperature at a given time they may contain changes in the water vapor and/or snow history, etc. Thus, comparison with GST series is useful for the ice core data also, since it can make the ice core reconstruction more precise.

The main goals of the work by Beltrami and Taylor (1995) were: (1) to calibrate the oxygen isotope data to the GSTs, and (2) to reconstruct GST history with higher resolution than that estimated from borehole data alone. Processing of the data included the estimation of the GST history from the borehole data, low-pass filtration of the oxygen isotope record to make it comparable with the calculated GST course, definition of the GST and filtered isotope data relation, and finally the reconstruction of the GST history from the calibrated oxygen isotope record. Figure 104 shows the oxygen isotope data and the reconstructed GST history. As shown, both datasets are generally coherent; however, high-frequency climate oscillations are lost in the GST diagram. Comparison of the GST with filtered oxygen isotope series on the 25-year scale of accumulation revealed stable linear relation between the two datasets. The oxygen isotope data S18O can be transformed into GST changes TG by the following simple formula:

With the help of additional proxy source, smoothed GST history in Figure 104, thus, can be reconstructed with the 25-year resolution.

Similar two-source reconstruction was performed by Harris and Chapman (2005) using tree ring and borehole GST histories. Except for the above-mentioned shortcomings of the tree-ring analysis, these data possess high short-term resolution, however,

Fig. 104. Top: Oxygen isotope data (25 years' averages) for the Agassiz ice cap core (Ellesmere Island, NWT, Canada) (data by Fisher et al., 1995; www.ncdc.noaa.gov/paleo/icecore/polar/ agassiz/data.html). Bottom: Ground surface GST history for borehole Neil (80.74°N, 83.08°W). (Drawn from the data by Beltrami and Taylor (1995).)

Fig. 104. Top: Oxygen isotope data (25 years' averages) for the Agassiz ice cap core (Ellesmere Island, NWT, Canada) (data by Fisher et al., 1995; www.ncdc.noaa.gov/paleo/icecore/polar/ agassiz/data.html). Bottom: Ground surface GST history for borehole Neil (80.74°N, 83.08°W). (Drawn from the data by Beltrami and Taylor (1995).)

have a strong detrending effect, and thus practically do not reflect low-frequency temperature variations. Borehole GST histories have opposite properties and may serve as a good complement to the tree-ring data because they are capable of capturing low-frequency events and reflecting real all-year temperature change and do not need calibration. The last but not the least is that both sources cover generally similar geographic areas. On the basis of available borehole temperature logs, Harris and Chapman (2005) have calculated an averaged reduced temperature-depth profile for land areas north of the latitude 20°N. The GST reconstruction performed using this averaged T-z profile has yielded a warming of 1.1K over the past 500 years. The tree-ring analysis from the same area (Briffa et al., 1992, 2001; see Figure 10, Chapter 1) has indicated considerably less warming for the same period. Merging of both data sources gave high-resolution surface temperature for the past 500 years, which is consistent with both borehole and tree-ring reconstructions without loss of the low-frequency variance. Comparison of borehole data including annual temperature variations with tree-ring reconstruction that reflects mainly warm-season conditions also provided an estimate of the long-term cold season (October-March) temperature variations. It has been demonstrated that continental extratropical Northern Hemisphere annual and cold season temperatures exhibited warming of 0.2 ± 0.1 K and 0.4 ± 0.3K, respectively, between 1500 and 1856 prior to the start of instrumental SAT record.

Valuable and probably the clearest comparison can be performed between borehole temperature reconstructions and SAT simply because both datasets are direct temperature measurements and do not need any calibration. Ground temperatures are directly related to the surface temperature forcing. Both sources incorporate climatic variations of all seasons of the year and thus can be interpreted as indicators of real annual temperatures. On the other hand, ground acts as a low-frequency filter that removes information about high-frequency climate variations from subsurface temperature-depth profiles. As a result reconstructed GST histories appear as smooth curves with resolution decreasing into the past. One of the first empirical relationships between annual mean GST and SAT has been presented by Kukkonen (1987) for the territory of Finland. It was based on the combination of air and ground temperatures measured on the meteorological stations all over the country and borehole temperatures extrapolated to the surface:

where TG and TA (°C) are annual mean ground and air temperatures, respectively. As seen on the annual scale the ground is warmer than the air. The ground temperature fluctuations are approximately 30% attenuated with respect to the air temperature. The fact that generally the mean annual SAT is lower than the corresponding GST was corroborated by numerous later measurements (see Section 2.5, Chapter 2). Because of different frequency content of the GST and SAT signals, a direct comparison of the detected climatic trends and/or the use of the above-described simple regression techniques may inspire erroneous conclusions (Harris and Gosnold, 1999). The estimation of POM described in Section 2.5 (Chapter 2) represents a more effective way to compile these two kinds of climatic information.

The idea of the joint processing of the measured temperature logs and the SAT was inspired by the above-mentioned complementary nature of the GST and SAT. To estimate the magnitude of recent temperature change, especially the amount of the recent global warming, paleoclimate reconstruction from the temperature-depth records can be suitably completed with a long-term meteorological SAT series monitored at weather stations. This idea was introduced by Harris and Chapman (1995, 1997) and provided a useful tool to assess the POM representing the temperature conditions that existed before the routine instrumental observations actually started some 100-250 years ago, i.e., the value against which the twentieth century climate warming could be referenced. Coupling the borehole temperature logs with the SAT series provides a more realistic benchmark than the models based on the inverted borehole data themselves.

Harris and Chapman (2001) inferred mid-latitude (30-60°N) climatic warming combining borehole temperatures with SAT series. For their reconstructions, authors have used temperature logs collected in the Global Borehole Temperature Database (www.geo.lsa.umich.edu/~climate). The measurement dates of the 439 temperature logs processed in this study ranged between 1958 and 1995. The curvature of the profiles was quite variable characterizing both recent warming and cooling similar to an example shown in Figure 108 (see below) for Canadian data. Observed variations have included both the natural climatic excursions as well as the environmental effects. A composite reduced temperature-depth profile was calculated by averaging all 439 T-z profiles forward continued to the year 1995. A positive curvature of the resulting profile hints the recent surface warming from the definite long-term mean level. A synthetic transient SAT series was computed by averaging the gridded meteorological data (Jones et al., 2000). The global SAT series from the same database is shown in Figure 4 (Chapter 1). Mean annual temperatures embrace the period 1856-2005. SATs are given as the temperature anomalies in accordance with the base period 1961-1990. For averaging, the authors used only the grid boxes that coincided with borehole locations. Calculated in this manner averaged SAT series was used as a forcing function for the POM estimate. The POM obtained by Harris and Chapman (2001) equals 0.71 K. Such result means that average GST during the long period prior to 1856 was approximately 0.7K below the 1961-1990 SAT mean. Reconstruction by Huang et al. (2000) based exclusively on the borehole data (see previous section) gives 0.9 K difference (warming) between the 1500-1700 mean temperature and the 1961-1990 reference level, which is very similar to the above POM estimate. According to Harris and Chapman (2001), the combination of the SAT and borehole data to obtain POM value appears to provide better fit to the data than the GST reconstruction alone.

Achieved in the recent years abundance of the global databases of high-resolution (annual and/or seasonal) proxy data, together with a few long instrumental and historical climate records available during the past few centuries, gives a possibility to reconstruct spatial patterns of temperature variations several centuries back in time. Reconstructions of the long-scale global/hemispheric trends can place the instrumental observations of the climate variations during the twentieth century in a longer term perspective, and thus provide a more reliable evidence of the role of potential climate forcings over time. The reconstruction of the Earth's temperature history for the past millennium using a variety of proxy records together with the borehole GST histories is generally connected to the often cited works by Mann et al. (1998, 1999) and Huang (2004) (see also collection of paleoclimate reconstructions presented on the web site by NOAA Satellite and Information Service and National Climatic Data Center, USA; www.ncdc.noaa.gov/paleo).

The main goals of the merging different paleoclimatic records are: (1) to obtain better spatial coverage and thus more reliable large-scale estimates, (2) to achieve better temporal resolution, and possibly the last but not the least (3) to verify the reconstructed climatic trends through the use of various independent sources of information. Borehole GST histories represent one of the best sources for the reconstruction of centennial trends. Century-long trends are also well presented in such proxies as the recession of glaciers. Detection of the log-term trends from, e.g., tree-ring data is difficult because of the above-mentioned detrending effect of the tree growth. Decadal/annual/seasonal resolution is provided by tree rings, ice cores, varved sediments, etc. Coupling of different kinds of climatic information can overcome the limitations of the individual proxies and provides useful comparison and validation of individual reconstructions. Combined with instrumental measurements and/or with documentary information, all these sources can significantly contribute to the gaining of reliable climatic histories on the global and/or hemispheric scales.

The large-scale reconstruction by Mann et al. (1999) was based on the tree ring, ice core, corals, and historical data, while described in the previous sections reconstructions by Pollack et al. (1998), Huang et al. (2000), Mann et al. (2003), and Pollack and

Fig. 105. Top: Annual scale temperature reconstruction for the Northern Hemisphere by Mann et al. (1999) for the period 1000 A.D. to 1980. Thick line represents its 50-year running average. Bottom: Integrated temperature reconstruction by Huang (2004). Thick line corresponds to the optimal surface temperature reconstruction by Mann et al. (2003) that uses terrestrial borehole data only. Temperatures are shown as the anomalies according to the 1961-1980 reference mean.

Fig. 105. Top: Annual scale temperature reconstruction for the Northern Hemisphere by Mann et al. (1999) for the period 1000 A.D. to 1980. Thick line represents its 50-year running average. Bottom: Integrated temperature reconstruction by Huang (2004). Thick line corresponds to the optimal surface temperature reconstruction by Mann et al. (2003) that uses terrestrial borehole data only. Temperatures are shown as the anomalies according to the 1961-1980 reference mean.

Smerdon (2004) have exclusively used borehole information. An inclusion of the GST reconstructions into the multiproxy and instrumental network was performed in the recent work by Huang (2004). To construct reliable Northern Hemisphere climate change history this research has merged three independent databases, namely 696 borehole GST reconstructions (significant part of them were used in the work by Huang et al., 2000), the twentieth century meteorological record, and the high-resolution multiproxy model by Mann et al. (1999) (Figure 105, top). This latter reconstruction is possibly the most often cited one. It is based on the tree ring, ice core, corals, and historical records of climate and shows the temperature variations in the Northern Hemisphere over the past millennium (from 1000 A.D. to 1980) with the annual resolution. According to the calculations by Huang (2004), borehole data alone suggest a cumulative warming of 0.9 K from the beginning of the sixteenth century to 1980. Obtained temperature increase is consistent with that estimated in the work by Huang et al. (2000) from smaller number of the borehole sites. The range of temperature variations in the record by Mann et al. (1999) does not exceed 0.7K. It exhibits variable, but generally decreasing trend up to the beginning of the twentieth century and the "hockey stick" shape in the last century (see Section 1.2.3 and Figure 11, Chapter 1). Smaller amplitude of the century-scale variability in comparison with that predicted by the geothermal reconstructions (see previous section) hints that this technique reflects well the short-term (annual) oscillations but cannot capture the long-scale trends. Recent studies with the GCM2 also suggest that centennial variations may have been larger (Bauer et al., 2003; Jones and Mann, 2004; von Storch et al., 2004; Zorita et al., 2004; Moberg et al., 2005).

The creation of the integrated surface temperature reconstruction consisted of the following main steps (Huang, 2004):

1. Creation of the extended surface temperature history by combining individual borehole estimates and 1900-1980 SAT record. For this purpose all available 696 individual GST reconstructions were gridded and combined with the twentieth century SAT temperatures gridded at the 5° X 5° area weighting base to prolong the latter database into the past. Data were adjusted according to the 1961-1980 reference period mean and averaged. Obtained SAT-GST surface temperature history is quite similar to those presented in Figure 100.

2. Generalized subsurface temperature anomaly-depth profile was calculated on the base of 1-D pure conductive approach using above extended surface temperature history as the surface boundary conditions. This profile was then used for compiled inversion with the mutiproxy time series (Figure 105, top).

3. As described in Chapter 2, the ability to incorporate additional information is one of the major advantages of both SVD and FSI inversion techniques. The merging of the generalized subsurface T-z profile with the multiproxy reconstruction by Mann et al. (1999) was performed by means of the functional space Bayesian inversion (FSI) (Shen and Beck, 1991) of this signal using annually resolved multiproxy series as the a priori model (see Section 2.3.5). The integrated time series of the temperature change over the last millennium represents the temporal domain of the inversion that should be estimated. It was parameterized3 at annual intervals to ensure readily the incorporation of the annually resolved multiproxy record as an a priori model.

Five-century long climate change series is presented in Figure 105 (bottom). The figure shows significant differences from the a priori model. On contrary to the multiproxy series in Figure 105 (top), it exhibits clearly colder past than the climate history by Mann et al. (1998, 1999) and the warming trend that could not be obscured by the high-frequency temperature fluctuations on either the annual or the decadal scales. The period from approximately 1575 to the beginning of the seventeenth century was the coldest during the whole reconstructed time interval, while the twentieth century was far warmer of the past five centuries. As shown, the integrated reconstruction by Huang (2004) coincides well with the optimal surface temperature history by Mann et al. (2003) calculated from

2General Circulation means the large scale coupled atmosphere-ocean motions arising as a consequence of differential heating on a rotating Earth. They restore the energy balance of the system through the transport of momentum and heat.

3Parametrization means the technique of the representation such processes that cannot be explicitly resolved by the model on the temporal (or spatial) scale. Parametrization provides the relationship between definite sub-grid and the larger scale for the investigated process (for details see Sections 2.3.4 and 2.3.5 of the Chapter 2).

the borehole data alone (Figure 105, bottom), as well as with the results of earlier reconstructions described in the previous section. Similarly to the mentioned hemispheric/global scale GST reconstructions, the warming of the twentieth century on the multiproxy results by Huang (2004) is seen more clearly as a simple continuation of recovery from the cold conditions prevailing in the sixteenth century that started far before the onset of the industrialization. On the other hand, warming rate appears to be accelerated in the nineteenth to twentieth centuries. This result corroborates the human-induced forcing of the natural climatic variations.

The primary limitations of the large-scale spatial/temporal proxy-based reconstruction arise from an increasingly sparse nature of the available proxy databases back in time. These databases can be extended only through merging of as large as possible results of paleoclimatic studies from different scientific branches. As demonstrated by the above example, the integrated reconstruction, treating together various paleoclimatic data (including borehole climatology), can display information that is weak and/or imperceptible in each record alone. Such integration of available climate histories and the search of common climate signatures in different reconstructions can provide more certain and broad conclusions about global mean temperature changes. So the integration of various paleoclimate reconstructions is more powerful than that either community carries out alone. The improved knowledge of the climate record can help to evaluate less speculative climate trends. Such improvements will lead to further advances in our empirical understanding of climate variations in the past millennium, and will allow for more meaningful comparison with the results obtained from model simulations of the past climate variation and empirical climate variability.

3.4 Is There Any Anthropogenic Component in the Present-Day Global Warming? Evidence From the Underground

One of the utilities of the climate reconstructions is their contribution to identifying the causes of climate change. In the above sections we have cited numerous examples (also including borehole climatology) that the Earth's surface has undergone unprecedented temperature increase over the last century. This climatic trend is known as the global warming. Global temperature has increased by about 0.6K (±0.2K) since the late nineteenth century. Even with a few interruptions this warming increases continuously. The most recent warming is particularly noteworthy because the rate of temperature increase is enormously high. SAT has increased about 0.2-0.3 K over the past 25 years (the period with the most credible data). Curiously, every single year since 1992 is in the current list of the 20 warmest years on record (http://www.data.giss.nasa.gov/gistemp/2005). Observed warming has not been globally uniform. The recent warmth has been greatest over North America and Eurasia between 40 and 70°N. Some areas (e.g., parts of the southeastern USA) have, in fact, cooled over the last century. The natural patterns of climate have been altered.

The anthropogenic origin of the global warming was proposed in 1938 by the English scientist G.S. Callendar (so-called "Callendar effect"). In spite of initial skepticism of scientific community, there is new and stronger evidence that most of the warming over the last decades may be attributable to human activities. Scientists know for certain that human activities are rapidly changing the composition of the Earth's atmosphere.

Increasing levels of GHG, like carbon dioxide (CO2), in the atmosphere have been well documented. However, it is not easy to detect to what extent the human-induced accumulation of GHG since pre-industrial times is responsible for the global temperature increase. Like many pioneering fields of the scientific research, the current state of global warming science cannot always provide indisputable answers to our questions.

3.4.1 Background

Global warming is a term used to describe the contemporary increase in the average temperature of the Earth's atmosphere and oceans. Although local temperatures fluctuate naturally, the rising trend that is known as "global warming" can be separated from historical and/or pre-historical climate variations that have occurred naturally. The term "global warming" is usually applied exclusively to the observed rapid global temperature rise during the last 100-150 years, which is believed to be related to an anthropogenic enhancement of the greenhouse effect. It appears that over the past 50 years the average global temperature has increased more rapidly than during the whole recorded paleocli-mate history. Linear trends can vary significantly depending on the period over which they are computed. However, the general prediction of the strong and rapid global climatic change seems to bear large level of significance.

The debates on the available scientific evidence of the global warming began in the second part of the twentieth century, almost immediately after the discovery of the unusual rising trends in the SAT series. Early chaotic discussions of the observed phenomenon oscillated widely from the hope that most predictions are wrong to the development of various models to explain the possible reasons for the climate change and generate its future scenarios. The Intergovernmental Panel on Climate Change (IPCC; www.ipcc.ch) played an important role in the study of the recent climate change and its consequences. It has been established by the World Meteorological Organization (WMO; www.wmo.ch) and by the United Nations Environment Program (UNEP; www.unep.org) in 1988 to assess scientific, technical, and socio-economic information relevant for the understanding of climate change, its potential impacts, and options for adaptation and mitigation. From 1990, the IPCC has published a series of technical papers and reports comprising the investigations that form standard basis for the reference widely used by scientists as well as by different experts and the policymakers up to the Kyoto Protocol.4 In its third assessment report "Climate Change 2001" (www.grida.no/climate/ipcc_tar) the IPCC affirmed that most of the recent 50-year warming can likely be attributed to the anthropogenic emissions of the GHG and suggested the need for actions to weaken this activity. It was shown that while the past climate history can be explained in the frames of a few well-known natural processes, such as solar variability and volcanic eruptions, none of these processes can be involved in the explanation of the recent dramatic warming. Thus, the modeling of human-created warming mechanisms seems to be indispensable to reproduce the modern temperature history.

4Kyoto Protocol to the United Nations represents the framework convention on climate change. It was adopted in 1997 in Kyoto (Japan) at the Third Session of the Conference of the Parties to the United Nations. Countries signed this convention agree to reduce their anthropogenic greenhouse gas emission by at least 5% below 1990 level during the 2008-2012 period. The Kyoto Protocol has entered into force in February 2005.

228 Borehole Climatology: A New Method on How to Reconstruct Climate 3.4.2 Greenhouse gases and climate change

For about the last decade, there has been an ongoing debate on the contribution of human activities to the global warming of the past century and especially on how anthropogenic activity will contribute to further warming that may occur during the twenty-first century. What is the physical basis for the fear of human-induced changes?

Similarly to other living organisms in all epochs, the mankind has influenced surrounding environment. However, an impact of human activities has drastically increased after the Industrial Revolution that began in the mid-eighteenth century in the UK and at the present time embraces the continental and/or global scales. The industrial revolution began with the invention of the steam engine. The most important human activities at present that may have an impact on both regional and global climate are connected with:

1. the combustion of fossil fuels and the biomass burning that produce GHG,

2. the emission of chlorofluorocarbons (CFC) and other halogen compounds that not only are strong GHG, but also play an important role in the depletion of the stratospheric ozone layer,

3. the emission of aerosols (propellants used in aerosol sprays) that affect the composition of the atmosphere, and

4. the change due to urbanization, agricultural practices, and forestry that influence the physical properties of the Earth's surface.

While the effect of other above-mentioned activities is complex and not yet well known, a detailed discussion on the influence of enhanced GHG content was performed in numerous research works.

Many chemical compounds found in the Earth's atmosphere act as GHG. GHG allow sunlight to enter the atmosphere. However, when it strikes the Earth's surface, some portion of this energy is re-radiated back as infrared radiation (heat). GHG absorb part of this energy, while the remainder escapes back into space. Detaining infrared heat energy they raise the temperature of the lower atmosphere and the Earth's surface in contact with it. Thus, the role of the GHG is that they trap the heat in the surface-troposphere system. If the atmosphere accumulated all the trapped heat, then the Earth's temperature would just rise and rise, but it is not the case. Over time, the energy amounts sent from the Sun and radiated back into the atmosphere became roughly equal, thus maintaining the temperature of the Earth's surface roughly constant at 14-16°C on average. It is the natural greenhouse effect that keeps the Earth's surface much warmer than it would be if there is no atmosphere.

The equilibrium is preserved till the amount of GHG in the air remains the same as well as the amount of heat arriving from the sun is constant. An increase in the concentration of the GHG gives a parallel increase of the infrared opacity of the atmosphere. This so-called radiative forcing causes an imbalance that can be compensated by the corresponding increase of the surface-troposphere system temperature. This is enhanced greenhouse effect. During the last 200 years, mankind has been releasing substantial quantities of extra-GHG into the atmosphere. These extra GHG are trapping more heat; thus, it is expected that the observed average annual SAT increase by about 0.6 K since

Table 9. Concentration growth of the important atmospheric greenhouse gases (except water vapor)

Greenhouse

Pre-industrial

Natural

Human

Total

% of total

gases

levela

additions

additions

concentration

Was this article helpful?

0 0

Post a comment