Info

Figure 1.4. Ice core data: sulfate record from the NGRIP core (Greenland) over the interval from ~ 10 to ~ 110 ka. The sulfate content was determined by continuously melting the ice core along its axis and measuring SO4 of the melt water by means of a photometer (continuous flow analysis, CFA; see Rothlisberger et al. (2000) and Bigler et al. (2002)); ppbw, parts per billion by weight. Meltspeed and signal dispersion limit the length resolution to ~ 1 cm over the measured record length (1530 m). In the young part of the record (t < 105 ka), the NGRIP timescale was obtained by tuning to the ss09sea timescale of the Greenland GRIP ice core (Johnsen et al. 2001) using the records of ice isotopes (North Greenland Ice Core Project members 2004), electrical conductivity and dielectric properties. In the old part, the NGRIP timescale was obtained by tuning to the GT4 timescale of the Vostok ice core (Fig. 1.3) using the records of ¿18O and methane concentration. (An absolutely dated alternative to the GRIP ss09sea timescale was published by Shackleton et al. (2004).) The sulfate record was finally averaged to 1-year resolution. Using the Ca and Na records, proxies for mineral dust and seasalt content, respectively, it is possible to remove peaks in the sulfate record from dust and salt input—the remaining peaks in the "excess" SO4 record, shown here, likely reflect the input from volcanic eruptions via the atmosphere. The record therefore bears the possibility to reconstruct volcanic activity throughout the last glacial period. (Data from Bigler M 2004, personal communication.)

100,000

sent those of the total southern hemisphere. However, such uncertainties are often unavoidable when general statements about the climate system are sought. All individual noise influences on a climate variable (natural variability, proxy and measurement noise) seem to produce a process Xnoise(T) with a PDF that is better described by a product than a sum of individual PDFs and that likely has a right-skewed shape, such as the lognormal distribution (Aitchison and Brown 1957).

1.3 Persistence

The other property of Xnoise(T) besides distributional shape regards serial dependence. The autocovariance, E [Xnoise(T1) ■ Xnoise(T2)] for

800-

_q

600-

-

400-

a

-

O

200-

0-

1 1x106-

-

-u c

-u

5x105-

n o

-

o

w

-

^

-

Q

0-

5-i

o

4-

-

(

3-

>

2-

o

-

1-

d

o

0-

O

600-i

£ 400-

z 200-

0-

Figure 1.5. Ice core data: Ca concentration (a), dust content (b), electrical conductivity (c) and Na concentration (d) from the NGRIP core (Greenland) during the onset of Dansgaard-Oeschger (D-O) event 5. The four variables were measured using CFA on the melted water (Fig. 1.4). ppb, parts per billion; ml-1, number of particles per ml; Sm-1, SI unit for electrical conductivity. A data gap (hiatus) exists at around 32,550 a in the dust-content record. Records were "downsampled" to annual resolution. The Ca record indicates variations of mineral dust transported to the atmosphere over Greenland, the dust content indicates atmospheric dust load, electrical conductivity is a proxy for input of soluble material (integrating various environmental signals) and Na is a proxy for seasalt. One climatological question is whether the changes in all four variables happened simultaneously at the onset of D-O event 5. D-O events are short-term warmings during the last glacial period. (Data from Rothlisberger R 2004, personal communication.)

32,100 32,300

32,700 32,900

Figure 1.5. Ice core data: Ca concentration (a), dust content (b), electrical conductivity (c) and Na concentration (d) from the NGRIP core (Greenland) during the onset of Dansgaard-Oeschger (D-O) event 5. The four variables were measured using CFA on the melted water (Fig. 1.4). ppb, parts per billion; ml-1, number of particles per ml; Sm-1, SI unit for electrical conductivity. A data gap (hiatus) exists at around 32,550 a in the dust-content record. Records were "downsampled" to annual resolution. The Ca record indicates variations of mineral dust transported to the atmosphere over Greenland, the dust content indicates atmospheric dust load, electrical conductivity is a proxy for input of soluble material (integrating various environmental signals) and Na is a proxy for seasalt. One climatological question is whether the changes in all four variables happened simultaneously at the onset of D-O event 5. D-O events are short-term warmings during the last glacial period. (Data from Rothlisberger R 2004, personal communication.)

200-1 150100-

0 0

Post a comment