Ground Temperature Histories Evidence of Changing Climate

Over the 100-200 years' long instrumental period, climate reconstruction leans against the amount and quality of proxy records and the skill of different techniques to infer climatic information from indirect measurements. Temperature-depth profiles measured in boreholes represent valuable source for paleoclimatic reconstruction that during the last two or three decades significantly contributed to the knowledge of the temperature variations from centennial to millennial timescales. Since sporadic ground surface temperature (GST) reconstructions in the 1970s and the 1980s, a vast amount of borehole temperature log analyses have been performed all over the globe. Because of high scatter in the data quality and different methods used for the GST reconstruction, a heap of obtained results is rather heterogeneous. However, in the regions with the dense coverage and a high quality of T-z profiles, common GST trends are clearly visible in the reconstructions of practically all researchers.

3.1 Timescales of the Reconstructed GST Histories (From Ice Age to the Present)

Provided that diffusion is a dominant heat transport process in the Earth, low-diffusivity crustal rocks have a considerable "memory" for the temporal variations of the GST. Fingerprints of the past climatic changes have been found in the temperature-depth profiles measured all over the world in boreholes from the Arctic to the tropic regions. Various methods have been developed in order to extract GST histories from these temperature logs for time intervals ranging from several decades to thousands of years. First attempts to decipher certain information on the GST changes from underground temperatures dates back to the early 1970s, and the corresponding "geothermal" method became generally known in the mid-1980s. The first specialized topical meetings were organized from the early 1990s, when the "borehole method" to reconstruct the past climate changes by the inversion of the borehole temperatures was recognized as a tantamount tool of paleoclimatic studies. Researchers from many countries are now routinely using this method, and a sizeable number of results have been published as well as regular meetings have been organized worldwide.

The first compilation of the studies inferring past climatic changes from underground temperatures has appeared in 1992 (Lewis, 1992). Different methods for the GST inversions were described and applied to the series of regional data. Most of the data came from North America. Three teams from Canada provided the bulk of the results. Also presented were paleoclimatic inversions from Western Utah and Alaska (USA). A few GST histories were inferred from temperature logs measured in Europe (Central France, Germany, and the Czech Republic). Two studies were devoted to the paleoclimatic reconstructions from warmer latitudes in Zaire (Central Africa) and Cuba. Most of the GST histories were relatively short and embraced period of not more than one millennium or so. Sparseness and sometimes clear occasional choice of the temperature logs for the inversions could not permit revelation of larger scale spatial and/or temporal patterns of the GST changes. Only climatic variations in Canada were relatively well documented. On the other hand, results presented in the issue by Lewis (1992) clearly demonstrated that the geothermal method, when used carefully, can provide further voluminous regional data on the paleoclimate change. Investigations have also shown that recent warming over approximately the last century derived earlier from meteorological surface air temperature (SAT) measurements is also seen in the GST histories in many regions. In central and eastern Canada this warming follows strong cold period, the initial warming being a return to the temperature averaged over the last few 100 years. Some of the studies detected good agreement of the past GST changes with the corresponding SATs and with the reconstructions using proxy data. All these works contributed significantly to the field of "Borehole Climatology" research, advancing the application of borehole temperatures to infer GST histories.

The reconstruction of the GST histories has drawn increasing attention under several international projects in the 1990s. The project No. 428 "Borehole and Climate" under the UNESCO International Geological Correlation Program was probably the most important of them. The project "Past climate change inferred from the analyses of the underground temperature field: Borehole temperatures and climate reconstructions" was proposed for the period 1998-2002 in order to collect suitable borehole temperature logs from different parts of the world, because their integrated analysis could significantly contribute to better quantitative assessment of the past climate changes and eventually extract the likely man-made (industrial) component of the recent warming from the natural long-term climatic variability. Practically all European countries (including Albania), USA, and Canada were involved in the project, followed by Brazil, India, China, Vietnam, Japan, Morocco, and Egypt. Considerable progress was made especially in Russia, Ukraine and Byelorussia, Romania, and Italy. The general scientific achievement of the project was the compilation of the fully operational database of boreholes with temperature logs and their corresponding GST reconstructions (

After 1992 the next collection of the borehole climate reconstructions from a number of regions all over the world was compiled by Beltrami and Harris (2001) and generally represented the results of the project No. 428. Published approximately eight years since the first special volume this issue presented wide new collection of works that captured well current directions of research on the Borehole Climatology subject. Little bit early a preliminary analysis of the worldwide dataset for climatic inferences from geothermal data have been published (Huang et al., 1997, 2000; Pollack et al., 1998). By the beginning of the twenty-first century the researchers obtained in their disposal wide amount of borehole reconstructions suitable for further conclusions. Boreholes distributed over almost entire landmass of the continents possess immense potential for the new observations of climate signature in many "white spots" of the global paleoclimatic map. At present there are numerical works trying to identify major spatial/temporal patterns of the changes in the global GST dataset. At the outset of discussion of hemispheric and global reconstructions that have been compiled from the analysis of borehole GST inversions, it is useful to list some of the characteristic GST reconstructions. Thus, the rest of this section represents the review of some of the most well known of these attempts.

According to their temporal length, all GST histories can be divided into three main groups:

1. embracing the Holocene (0.01Ma) or so,

2. one or two millennia, and

3. the recent 100-200 years.

The first time interval will be described in detail in Section 3.5 with respect to the deep boreholes. In further subsections we present the summary of the 1000-2000 years' long GST temperature record as well as of the evidence of the recent warming. Most of existing boreholes are generally 500-600 m deep and thus contain information only about approximately the past 500 years. GST history from 1500 to the present is thus especially of interest, because these data represent the bulk of available GST reconstructions. Section 3.2 is devoted to this time interval.

3.1.1 GST changes in the last two millennia (spatial and temporal patterns)

Temperature measurements to the depth of 200-300m are the most common in geother-mics, while the data from only 700-1000m deep holes can reveal climate excursions of the past one to two millennia. Thus, the regions where the GST histories for such long timescales were inferred represent relatively small part of existing inversion results. Below we describe the most important areas where such kind of the GST reconstruction was performed.

The Czech Republic Probably the most dense borehole network exists at the relatively small territory (approximately 79000 km2) of the Czech Republic, landlocked country in Central Europe. There are over 200 temperature-depth profiles in this region measured between 1963 and 1992. Ninety-eight of them, which have been carefully selected (Figure 74), have been used to reconstruct the past GST conditions (Safanda and Kubik, 1992; Bodri and Cermak, 1995, 1997a, 1999; Safanda et al., 1997). Criteria for screening the data were rather severe (see references above). For example, to minimize the possibility of biasing the results by topography and groundwater circulation, the authors have excluded all boreholes located in the mountainous regions and/or displayed evidence of the water movement (see Section 2.7). As shown in Figure 74, the studied area has a typically uneven distribution of data. This clustering, with a few exceptions, is a

Fig. 74. Map of the Czech Republic showing borehole sites for which temperature logs were processed. (Note: Nearby holes are shown as single dots; triangles show locations of the meteorological stations.)

common feature in the borehole climatology. It is the product of the fact that most temperature loggings were performed in areas of intense industrial interests, such as mineral exploration, coal and oil prospection, and hydrogeological survey.

Borehole Holubov (HO-1) represents a typical example of the precise temperature logs and the GST reconstruction in this region of Europe. It is located in southern Bohemia, about 150km south of Prague. Borehole is situated in a flat area apparently free of topographic and/or groundwater flow disturbance. The high-quality temperature logging was performed in 1972 down to 700 m after the hole had been in equilibrium for several years. The hole penetrated the alternating layers of granulites and peridotites, typical rocks of the ultrabasic structures of the Bohemian Massif. Thermal conductivity and heat production were measured for numerous samples (Cermak, 1975). The mean rate of heat production is 0.63 ± 0.06 |iW/m3 (Cermak, 1975). For the purpose of the present analysis heat production of this rate has negligible effect on the inversion results. The temperature log has a small but expressive curvature ("U-shape") in its uppermost part indicating recent warming (Figure 75). The effect of the topography on the measured temperatures was assessed by Safanda and Kubik (1992) to be approximately —0.5 mK/m and relatively constant with depth. Figure 75 also presents the vertical variation of the thermal gradient and heat flow for borehole HO-1. The gradient is low in the uppermost part of the hole, then increases and reaches its "undisturbed" value (compare with synthetic example in Figure 20, Chapter 2).

The GST reconstruction is shown in Figure 76. About 1000-year long climatic history can be recovered from measured T—z profile. The characteristic time events deduced from the inversion indicate a warmer period with maximum between 1230 and 1320 A.D., followed by a colder period with a minimum at 1650-1700 A.D., then some warming before 1880,

Fig. 75. Left: Temperature log (circles) and thermal gradient for borehole Holubov, HO-1 (the Czech Republic) (Bodri and Cermak, 1995). Center: Temperature log on the reduced scale (circles) and thermal conductivity (squares). Right: Vertical variations of heat flow for borehole HO-1.
Fig. 76. Reconstructed GST history for hole Holubov (HO-1) (Bodri and Cermak, 1995). The "Medieval Warm Period" is centered near 1200 A.D. and the "Little Ice Age" near 1650 A.D.

cooling with a minimum at 1950 and pronounced recent warming after 1960. Prior to 1960 the calculated temperature oscillations did not exceed 1.5 K; however, the amplitude of the recent warming may have reached 1.7 K. The reconstructed long-term surface conditions correspond generally well to the Medieval Warm Period and the Little Ice Age.

Defining the temporal and spatial nature of climate changes is an important stage in understanding the underlying causes of climate variation. Similar GST histories were

Fig. 77. Spaghetti diagram: GST histories reconstructed for 98 holes in the territory of the Czech Republic.

derived for all 98 Czech boreholes and used to reconstruct both country-averaged and regional patterns of the respective climate changes. Figure 77 shows transient GST history of 98 boreholes in the Czech Republic. The display of only the transient components of GST enables an easier comparison of the climate change from the regions with different steady-state conditions. The GST histories shown in this figure represent a typical example of the "spaghetti diagrams" characteristic for graphical representation of mete-orological/climatologic data (see Shen et al., 1995; Figure 30, Chapter 2 and comments on the web site When all GST histories are drawn, the form usually looks like a bowl of spaghetti. Such diagrams not only are common in the borehole climatology (see, e.g., Figure 11, Chapter 1) but also usually show the results on an almost unintelligible scale. The effectiveness of the spaghetti diagrams is in how it illustrates an amount of possible GST change. In the case of the GST histories an ensemble of curves implies high spatial/temporal variability of the climate changes and/or could suggest that the effects of representational errors may not have been adequately removed (Shen et al., 1995; Gosnold et al., 1997; Cermak and Bodri, 2001). As was shown by Pasquale et al. (2005), the spaghetti diagram of the GST changes reconstructed from the temperature logs measured in the suite of boreholes in northern-central Italy clearly reflects sharp climatic differences between the Tyrrhenian and Adriatic sides of the Peninsula. The Italian orography and the position of the country surrounded by the Mediterranean Sea imply strong effects of the local air circulation that cause high spatial and temporal climate variability. Simple averaging of the GST histories provides highly smoothed regional curves that could not account for the general climate trend in the region.

Fig. 78. The diagram of occurrences of extremes of climatic events in the past two millennia with the reconstructed air temperatures (50-year running averages) in the Czech Lands plotted as deviation from the 1851 to 1950 mean (Brazdil, 1990).

It can be demonstrated, however, that an ample set of "spaghetti-curves" from extensive area may be confusing only at the first sight and certain general climatic trend can be revealed in almost all cases. One of the more advantageous methods to delineate climatic trends characteristic for the vast investigated area is to combine detected extremes. Thus, Figure 78 summarizes the occurrence times of the extremes of climatic events reconstructed for all 98 inverted GST histories. The times of temperature minima are shown as "cold" and occurrences of maxima as "warm". As one might expect, Figure 78 exhibits a much more definite pattern of climate changes than the "spaghetti diagram" above. Three early episodes can be distinguished: a cold period between the eighth and tenth centuries, the renewal of general warmth (Medieval Warm Period) with its culmination around 1250 ± 50 A.D., and the cold conditions (the Little Ice Age) with minimum at 1650 ± 30 A.D.

Figure 78 also presents the comparison of the reconstructed GST extreme series with the annual air temperature (50-year running averages) reconstructed for the Czech Lands by Brazdil (1990) showing as the deviation from the 1851 to the 1950 mean. Brazdil's data are based on instrumental observations, and were completed by written historical sources on various indirect indicators of climate, such as grape and hop harvests. The coherence of both series is quite high. The weak appearance of the Medieval Warm Period in the Brazdil's record may be due to the relative scarcity of the historical data available from this period. While the three early climatic episodes appear as well-marked clusters in the extreme occurrence diagram, the appearance of the Little Ice Age is more erratic and is dispersed over as long as 300 years. Such a feature of the Little Ice Age in the study area is also clearly visible in the climatic record by Brazdil (1990). The cold climate conditions at the territory of the Czech Republic dominated most of the sixteenth and eighteenth centuries, in some cases probably alternated with provisional returns of

Fig. 79. Comparison of the last 250-year segment of Figure 78 with the mean annual temperatures (10-year running means) for the Prague (Klementinum), Vienna, and Munich meteorological stations.

warmth. The earlier climax of this period around 1650 ± 30 A.D. was observed in the 24 GST histories from different locations, thus believed to be characteristic of most of the territory of the Czech Republic.

Since the beginning of the nineteenth century a general warming has dominated the climate pattern interrupted by several shorter, generally not more than decade-long fluctuations of relatively colder and warmer conditions. Using the time occurrence of the reconstructed GST extremes, the next characteristic times of minimum (min) and maximum (max) alternating extremes were found: max 1730 ± 20 A.D., min 1780 ± 10 A.D., max 1820 ± 10 A.D., min 1880 ± 10 A.D., max 1935 ± 7 A.D., min 1943 ± 5 A.D., and max 1976 ± 3 A.D. Figure 79 presents the comparison of the last 250-year segment of Figure 78 with the meteorological series of mean annual air temperatures recorded in Prague, Vienna, and Munich. It shows that the relatively warmer period of the early nineteenth century with a subsequent cooling, indicated by geothermal method, also appears in these records. All records then confirm that the 1880s were the coldest decade. The warmer decade of the 30s and a colder decade of the 40s of the twentieth century, which have been recovered by geothermal method, are confirmed in the instrumental records. On the other hand, certain setbacks of the warmer temperatures in the meteorological series in periods 1948-1953, 1966-1967, 1971-1977, 1981-1983, and 1988-1990, may be regarded as a high-frequency noise in the general warming trend of the twentieth century rather than individual climatic events. These episodes did not appear on the inverted temperature logs but were resolved as a single warming event with its climax around 1976.

The derived GST histories have also been used to construct a tentative regional pattern of climatic changes for several distinct climatic epochs on the territory of the Czech Republic. Previously existing information on climate excursions of the last millennium for this region was very poor and, like temperature series by Brazdil (1990) presented in Figure 78, was based mainly on the analysis of documentary sources and the instrumental mean annual air temperature records (since 1771) from the meteorological station Prague-Klementinum (Brazdil, 1990; Brazdil and Kotyza, 1995).

The climate changes detected by the GST reconstructions were mapped for the following periods: 1100-1300 A.D. (Little Climatic Optimum), 1400-1500 A.D., 1600-1700 A.D. (main phase of the Little Ice Age), and for the most recent climate trend after the year 1960. The rate of the temperature change calculated as derivative of the reconstructed GST history was used as a robust indicator of the climatic trend. As the span of the reconstructed GST history depends on the borehole depth, not all boreholes can cover all time intervals studied. The investigated territory was divided into a regular 1° X 1° latitude-longitude grid network, and the obtained values were averaged over all grid elements. This procedure enabled to overcome the influence of possible incoherency of GST histories that appeared in the "spaghetti diagram" (Figure 77) as well as the uneven distribution of the drillhole sites.

Regional pattern of the average warming rate between 1100 and 1300 A.D. is presented in Figure 80 (top). The existence of the Medieval Warm Period is clear; during this time the entire investigated territory experienced warming, with a maximum rate of up to 0.2K/100year, notably in Central Bohemia. A general turn toward colder climate came during 1200-1400 A.D.; the time from 1550 to 1700 A.D. is regarded as the main period of cold climate. However, as was described in Chapter 1 (Section 1.1) the onset of the Little Ice Age as well as its duration was by far not identical throughout Europe. According to the above-mentioned climatic reconstructions by Brazdil (1990), this period was quite erratic at the territory of the Czech Lands and continued at least to the end of the eighteenth century (or even extended into the nineteenth century). Dominant cooling in some cases alternated with provisional returns of warmth (e.g., between 1520 and 1560 A.D. and in the 1670s). To investigate the gradual development of the cold conditions, separate maps were constructed for two stages of this period: (1) 1400-1500 A.D. and (2) 1600-1700 A.D. (Figure 80). These maps illustrate the advance of the cold conditions from the east. While during 1400-1500 A.D. the westernmost regions of the investigated area still showed certain warming trend, possibly the relict of the Medieval Warm Period, between 1600 and 1700 A.D. the entire investigated territory was already subjected to massive cooling. The cooling rates in this period were at least three to five times larger than in the time interval of 1400-1500 A.D.

Collection by Lamb (1977) of the winter severity/mildness indices is a useful database for independent comparison of above GST reconstructions. This index represents the excess number per decade of cold winter months (December, January, and February) over months of opposite character. This index may vary between +30 and —30. Excess of cold months is counted negative. Lamb (1977) has gathered collection of indices from 1100 A.D to the 1960s for three positions across Europe near latitude 50°N. The first case embraces the summary of data for the UK (longitude approximately 0°E), the second one corresponds to Germany (approximately 12°E), and the last one is for Russia (approximately 35°E). Extreme decade values in this database range from —28 to 12 over the investigated period. Figure 81 shows the longitude-time map of winter severity index. The findings represented by Figure 80 are in good agreement with the index map. According to Figure 81, the period between 1400 and 1500 A.D. can be characterized

Fig. 80. Top: Regional pattern of the warming rate (K/100 years) between 1100 and 1300 A.D. in the territory of the Czech Republic. Center: Regional pattern of the climate change rate (K/100 years) between 1400 and 1500 A.D. Bottom: Regional pattern of the climate change rate (K/100 years) between 1600 and 1700 A.D.

Fig. 80. Top: Regional pattern of the warming rate (K/100 years) between 1100 and 1300 A.D. in the territory of the Czech Republic. Center: Regional pattern of the climate change rate (K/100 years) between 1400 and 1500 A.D. Bottom: Regional pattern of the climate change rate (K/100 years) between 1600 and 1700 A.D.

Fig. 81. Winter severity index in different European longitudes near 50°N from 1100 A.D. to 1969. Dark colors indicate colder winters.

by the predominance of cold winters with the clear eastward cooling trend. Extremely cold winters also prevailed all over the 1600-1700 A.D. period. The severity of winters in this period was higher than between 1400 and 1500 A.D. Winter severity rapidly increased to the east.

Even when the most borehole GST reconstructions in the Czech Republic, like that presented in Figure 76, have yielded strong ground warming for the whole twentieth century or so, the detailed GST history of the last few decades may be more complicated. The regional trend since approximately 1960 for this region was first reported by Safanda et al. (1995). The authors concluded for the last-century climate conditions: (1) an insignificant change or even cooling in Western Bohemia, (2) 0.2-0.5 K warming over the extensive southern and central parts including the capital Prague, and (3) 0.2-0.3 K cooling in the NE part of the Bohemian sedimentary basin following previous 150-200 years of almost 1K warming. Comparison of the obtained GST histories with combined air and soil meteorological records disclosed that revealed difference in regional characteristics may have its real regional background.

Fig. 82. Left: Tentative attempt to construct the regional pattern of the last 30-year warming rate (K/year) in the Czech Republic from GST reconstructions. Right: Map of the recent warming rates obtained from the meteorological mean annual air temperature records. Locations of the meteorological stations are shown in Figure 74.

Fig. 82. Left: Tentative attempt to construct the regional pattern of the last 30-year warming rate (K/year) in the Czech Republic from GST reconstructions. Right: Map of the recent warming rates obtained from the meteorological mean annual air temperature records. Locations of the meteorological stations are shown in Figure 74.

Figure 82 (left) presents the tentative regional map of the GST change rate since approximately 1960 (Bodri and Cermak, 1999). Warming has been most pronounced around Prague and vicinity and decreased to the south and southwest. Another area of significant warming is located in the easternmost part of the country. The western part of the Czech Republic even goes through an inexpressive cooling (-0.007 K/year). Five holes from the western group were investigated by Clauser and Mareshal (1995) and seven by Safanda et al. (1997) who also reported negative rate of the temperature change in the most recent past. To verify the obtained climate history, we used instrumental records of the mean annual air temperature for the period 1961-1996 at 30 local meteorological stations (Figure 74). For each station the linear trend was calculated; these trends then were used to compute contour lines. The resulting map of the "meteorological" warming is presented in Figure 82 (right), being generally of the same order and similar to that given by the "geothermal" data. The most significant warming was characteristic for the central and eastern parts of the Czech Republic, and gradually decreases to the south and north. Figure 83 shows the results of spatial correlation between both patterns. Spatial correlation was calculated from the dense network computed for the construction of contour lines using 3 X 3 moving window. As shown, both patterns show generally high correlation. An independent analysis of the SAT records from Czech meteorological stations for the same period has revealed warming trends that fall in the interval from 0 to 0.04 K/year with characteristic regional warming rate of 0.0283 K/year (Cermak et al., 2000). Approximately 60% of the data fall within 0.02-0.03K/year interval.

The reasons for the general warming of the last three decades can only be speculated upon. The global warming phenomenon should be mentioned among the possible candidates. On the other hand, the regional distribution of warming presented in Figure 82 gives a certain idea on a potential human impact on climate conditions. The highest indicated rates of the recent GST-warming correspond to the industrial and relatively densely inhabited regions of Prague, northern Bohemia, and Ostrava coal basins; the lowest rates are characteristic for the SW and S slopes of the Bohemian Massif, areas generally forested and less industrialized. An analysis of the spatial pattern of the recent SAT trends by Cermak et al. (2000) has confirmed conclusion by Bodri and Cermak (1999) that more

0 0

Post a comment