Isotope Dendroclimatology

Many studies have demonstrated empirically that variations in the isotopic content of tree rings (813C, 8lsO, and 82H) are in some way related to climate.33 Early studies used whole wood samples but some studies now suggest that latewood may need to be isolated from earlywood to get a clear signal of climatic conditions in the growth year (Switsur et al., 1995; Robertson et al., 1995). In mid- and high latitudes, there is a positive relationship (sensu lato) between the oxygen and hydrogen isotopic composition of rainfall and temperature (Rozanski et al., 1993) so it is reasonable to expect that the isotopic content of wood in trees might preserve a record of past temperature variations in such regions (Epstein et al., 1976). The problem is

33 To avoid the difficulties of chemical heterogeneity in wood samples, a single component, a-cellulose (polymerized glucose), is extracted for isotopic analysis; a-cellulose contains both carbon-bound and oxygen-bound (hydroxyl) hydrogen atoms, but the latter exchange readily within the plant. It is therefore necessary to remove all hydroxyl hydrogen atoms (by producing nitrated cellulose) to avoid problems of isotopic exchange after the initial period of biosynthesis. For a discussion of isotopes, deuterium/hydrogen (D/H) and 180/160 ratios, see Sections 5.2.1 and 5.2.2.

that additional isotopic fractionation (of hydrogen, oxygen, and carbon atoms) occurs within trees during the synthesis of woody material and these biological fractionations are themselves dependent on many factors, including temperature, relative humidity, and wind speed (évapotranspiration effects) (Burk and Stuiver, 1981; Edwards, 1993). Nevertheless, in spite of such complications, a number of studies have found very strong positive correlations between 813C, 8lsO, and 62H and temperature, and negative correlations with relative humidity. The exact temperature relationship varies, but in several studies of 813C and 8lsO, it is commonly in the range of 0.3-0.4%o per °C (Burk and Stuiver, 1981; Yapp and Epstein, 1982; Stuiver and Braziunas, 1987; Lipp et al., 1991; Switsur et al., 1995). Most studies have focused on only the last few decades, but a few have attempted longer paleo-climatic reconstructions. For example, based on the strong correlation observed in recent data between August temperature and 813C in fir (Abies alba) from the Black Forest of southern Germany, Lipp et al. (1991) reconstructed temperature back to A.D. 1000 (Fig. 10.31). This record suggests that there was a steady decline in temperature from the early fourteenth century to -1850, with an earlier warm episode centered on A.D. 1130. Other long-term temperature reconstructions include a 2000 yr 813C-based record from the western United States (Stuiver and Braziunas, 1987) and a 1500 yr 82H-based record from California (Epstein and Yapp, 1976). There is little similarity between these two records, possibly reflecting real temperature differences, but perhaps also highlighting the many complications that may confound any simple interpretation. Certainly, the processes involved are complex. For example, Lawrence and White (1984) found that 8D in trees from the northeastern United States does not contain a strong temperature signal, as might

L Â

i 1 i 1 i 1 i ÉlllK i

1,1,1,1,

1

1

3

2

O

o

CD

1

5

a>

0

E

a>

H

-1

0

a.

o

o

-2

V)

-3

1200

1400

1600

1800

2000

Year

FIGURE 10.31 Average 8I3C values of cellulose extracted from the latewood of 19 fir trees (Abies alba) in the Black Forest, from A.D. 1004-1980.Values shown are smoothed by a 100-yr Gaussian lowpass filter; values from 1850-1980 are corrected for the effects of contamination by fossil fuel C02. Based on calibration over recent decades, the scale of August mean temperature is shown on the right, relative to the long-term mean (Trimborn et al„ 1995).

be expected, but is inversely correlated with summer rainfall amount. By contrast, Ramesh et al. (1989) found 8D in teak from western India was positively correlated with rainfall. In each case it is possible to construct an explanation for the observed relationship, but one feels that this approach is a little too ad hoc. There is a real need to build on the empirical approach by constructing comprehensive models to help in understanding all the processes involved (Edwards and Fritz, 1986; Clague et al., 1992; Edwards, 1993). Isotopic dendroclimatology has much potential, but significant efforts are still needed to develop reliable paleoclimatic records to complement ring-width and densitometric studies.

10.4.1 Isotopic Studies of Subfossil Wood

Yapp and Epstein (1977) showed that 8D in wood was strongly correlated with 8D in associated environmental waters, which were consistently 20-22%o higher than the nitrated cellulose (Epstein et al., 1976; Epstein and Yapp, 1977). As 8D of annual precipitation is correlated (geographically) with mean annual temperature (Yapp and Epstein, 1982) long-term records of 8D from trees may be a useful proxy for temperature variations. This idea is supported by mapping 8D values from modern plants and comparing them with measured 8D values of meteoric waters (Fig. 10.32). It is clear

Meteoric Water Latitude

FIGURE 10.32 Isolines of 5D based on modern meteoric waters (i.e. in precipitation) compared to 8D values inferred from cellulose C-H hydrogen (cellulose nitrate) in modern plants (underlined values).With the exception of only two Sierra Nevada samples, the inferred values differ by an average of ~ 4%o from the values measured in precipitation samples (Yapp and Epstein, 1977).

FIGURE 10.32 Isolines of 5D based on modern meteoric waters (i.e. in precipitation) compared to 8D values inferred from cellulose C-H hydrogen (cellulose nitrate) in modern plants (underlined values).With the exception of only two Sierra Nevada samples, the inferred values differ by an average of ~ 4%o from the values measured in precipitation samples (Yapp and Epstein, 1977).

that the 8D in plants provides a good proxy measure of spatial variations in 8D of precipitation. Assuming that this relationship has not changed over time, it is possible to reconstruct former 8D values of meteoric water by the analysis of radiocarbon-dated subfossil wood samples (Yapp and Epstein, 1977). Figure 10.33 shows such a reconstruction, for "glacial age" wood (dated at 22,000-14,000 yr B.P.). Surprisingly, the ancient 8D values at all sites are consistently higher than modern values (an average of +19%o). The higher 8D of meteoric waters implies that temperatures over the ice-free area of North America were warmer in late Wisconsin times than today. However, there are many other factors that could account for the high 8D values observed. In particular, 8D values in the wood may reflect rainfall in warm growing seasons and so would be isotopically heavier than the annual values mapped in Fig. 10.32. This would be especially true if more of the extracted cellulose came from late-wood rather than earlywood.

Other factors which could help to explain this surprising set of data include: (a) a reduction in the temperature gradient between the ocean surface and the adjacent precipitation site on land; (b) a change in 8D of the ocean waters as a result of ice growth on land (probably corresponding to an increase in oceanic 8D of 4-9%o); (c) a change in the ratio of summer to winter precipitation; and (d) a positive shift in the average 8D value of oceanic water vapor, which at present is not generally evaporating in isotopic equilibrium with the oceans.

Pictures For Dendroclimatology

Wisconsin glacial maximum, as inferred from 8D values of tree cellulose C-H hydrogen (underlined).The approximate position of the southern margin of the ice sheet is shown at its maximum extent (hatched line).The "glacial age" meteoric waters of the 15 sites have, on average, 8D values which are l9%o more positive than the corresponding modern meteoric waters at those sites as deduced from the data shown in Fig. 10.32.The "glacial age" distribution pattern of 8D values is similar to the modern pattern, but is systematically shifted by the positive bias of the ancient waters.The North American coastline shown is that of today and does not take into account the lower sea level at the time of glacial maximum (Yapp and Epstein, 1977).

Wisconsin glacial maximum, as inferred from 8D values of tree cellulose C-H hydrogen (underlined).The approximate position of the southern margin of the ice sheet is shown at its maximum extent (hatched line).The "glacial age" meteoric waters of the 15 sites have, on average, 8D values which are l9%o more positive than the corresponding modern meteoric waters at those sites as deduced from the data shown in Fig. 10.32.The "glacial age" distribution pattern of 8D values is similar to the modern pattern, but is systematically shifted by the positive bias of the ancient waters.The North American coastline shown is that of today and does not take into account the lower sea level at the time of glacial maximum (Yapp and Epstein, 1977).

Whatever the reason, the isotopic composition of glacial age precipitation has important implications for the isotopic composition of the Laurentide Ice Sheet, which is generally assumed to have been composed of ice that was very depleted in lsO and deuterium. Using 8D values from trees growing along the shores of Glacial Lakes, Agassiz and Whittlesey, Yapp and Epstein (1977) calculated that the 8180 value of waters draining from the former Laurentide ice sheet probably averaged around -12 to -15%o, far higher than would be expected by analogy with glacial age ice in cores from major ice sheets. Measured 8D values from glacial age ice in cores from Greenland and Antarctica (Chapter 5) average 80%o below those of modern precipitation in the same area. These results are thus somewhat enigmatic and it would be extremely valuable to extend this work to other formerly glaciated areas (particularly Scandinavia) to study "glacial age" 8D values in more detail because they have an important bearing on the interpretation of other paleoclimatic records, particularly ice and ocean cores. On the other hand, there may have been other factors operating to alter the 8D/temperature relationship observed today, or to bias the wood isotopic signal in some way, illustrating once again the difficulty of interpreting isotopic signal in tree rings.

Was this article helpful?

0 0
Renewable Energy Eco Friendly

Renewable Energy Eco Friendly

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable.

Get My Free Ebook


Post a comment