Mongolia, with a largely agrarian economy, is greatly concerned about regional and global changes in climate. The country has an extreme continental climate comprised of very cold winters and hot summers. Much of the country is semi-arid to arid and precipitation variations have a profound impact on the economy.
The northwest part of the nation is within the region of the Mongolian or Asiatic high, one of the most extreme atmospheric pressure regions of the world. This high influences circulation for much of central Asia. The climate of Mongolia is dominated by advected air masses although in summer local radiation balances and convective storms also influence precipitation and temperatures.
Recorded climatic data are rather short with few meteorological or hydrological records extending back past the 1940s. Thus there is need for better and more extended climatic information for the country itself and for the region's role in larger-scale climatic variations.
In 1995 a Mongolian-American Tree-Ring Project (MATRIP) was started and field collection of samples took place in 1995 and 1997. A paper describing the first results came out in 1996 (Jacoby et al., 1996) [results here Figure 1] and a more intensive phase of the project began in 1997.
The three main species we have sampled and found useful are Siberian pine (Pinus sibiricaDu Tour), Scots pine (P. sylvestrisL.), and Siberian larch (Larix sibiricaLedebour).
We tried to find sites with the least evidence of human disturbance. There was evidence of fire in three of the sites where we sampled Scots pine but there seemed to be little effect in most of the cores and resulting ring-width series. With its thick bark, the mature Scots pine are relatively resistant to fire damage.
We have found living trees over 700 years in age and a radiocarbon analysis indicates an age of about 700 AD for the inner rings of a dead snag. There are abundant dead snags at a few sites. Thus there is potential for millennial length chronologies.
Analyses of recorded temperatures by Dagvadorj and Mijiddorj (1996) showed increases for fall, winter and spring temperatures but decrease in summer temperature for the 1940 to 1995 period (Figure 2).
Dagvadorj and Mijiddorj (1996) also note the following changes in the recorded temperatures: winter heating degree days are less, the growing season is longer by about 10 to 20 days due to the warmer spring and fall, the extreme heat in summer is less, and annual temperatures have increased by about 1.8 degrees C in western Mongolia, 1.0 degree C in central Mongolia, and 0.3 degrees C in eastern Mongolia.
The Tarvagatay tree ring-width index series shows good agreement with an Arctic zone temperature reconstruction based primarily on latitudinal tree-ring records (D'Arrigo & Jacoby, 1992; Jacoby et al., 1996).
The conclusion is that there is definitely an unusual warming for this site in Mongolia. This warming and previous low-frequency variations reflect temperature changes similar to the larger-scale, hemispheric temperature trends (Jacoby et al., 1996).
Changes in precipitation have occurred for central Mongolia. The recorded data from 1940 to present shows an increase in total annual precipitation. We divided the precipitation record into three approximately equal time periods for comparison.
The greatest monthly increase is in August, the second wettest month of the year, when the precipitation has increased by about 33% (Figure 3). For May, a much drier month, precipitation has decreased by about 25%. Ring-width chronologies from two moisture-sensitive, lower forest border sites were used to make reconstructions of precipitation. For each reconstruction several meteorological stations were merged to produce a better regional precipitation record for the area around the site.
The tree-ring data from Urgun Nars (at 48 34.6'N, 110 32.7'E, elevation 1070 m) in east central Mongolia had the highest variance explained by monthly precipitation data. The sampled species is Scots pine.
Using the data from January of the prior year through October of the current or growth year, precipitation could explain 58% of the tree-ring variation using the 1942-1995 period. The standard chronology from ARSTAN was used in the modeling. A positive response to mid summer to fall precipitation of the prior year and to spring and summer of the current year was shown (Figure 4).
Two models were made based on this result: (1) summer precipitation of June, July, and August and (2) annual precipitation for a year extending from prior August to current July. Each model included ring-width indices for the current year and for the following year because the correlations (Figure 4) indicated the effect of prior year precipitation on the next year's growth.
The models explained 42% of the variation in summer precipitation and 54% of the variation in annual precipitation. Due to the shortness of the meteorological record, calibration-verification analyses were not performed. The reconstructions are shown in Figure 5.
The other moisture-stress tree-ring data are Siberian pine samples from Manzshir Hiid (47 45.96'N, 106 59.72'E, Elev. 1740-1770m) in central Mongolia. The variation in tree growth explained by monthly precipitation was 41%. The standard chronology was used and the correlations with monthly precipitation are shown in Figure 6.
The reconstruction is for February through July and also uses current year and following year ring-width indices for estimating the seasonalized precipitation. The model explains 36% of the precipitation variation (Figure 7).
Both reconstructions are in general agreement during the 1900s (Figure 8). There are some differences, especially in the mid to late 1800s. We plan to explore these differences to understand the current changes in spring and summer precipitation distribution.
Neither reconstruction shows severe drought in the 1820s, especially 1827, that is shown in the historical reconstruction (Mijiddorj & Namhay 1993). The reconstructions are in agreement in the 1920s when both tree-ring reconstructions and the historical reconstruction all show severe drought.
The tree-ring reconstructions are for central and east central Mongolia whereas the historical reconstruction is for all Mongolia therefore some of the differences may be due to different geographical coverage. More tree-ring reconstructions are needed before full comparisons can be made.
Some earlier studies indicated quasi-solar periodicities in tree-ring data (e. g. Lovelius et al. 1992) and drought (Mijiddorj & Namhay 1993). Our analyses indicate that periodicities are very dependent on the tree response to climate.
Temperature sensitive trees as represented by the Tarvagatay site trees have very little periodicity. The trees from moisture sensitive sites show much more spectral power indicating periodicities in the 22-year range and 3 to 4-year ranges.
The 22-year periodicities in moisture-stressed trees and in drought occur in analyses of tree-ring data and drought in the coterminous United States of America (Cook et al. 1997) and were previously hypothesized by A. E. Douglass (1919). The 3 to 4-year periodicities are in the range of ENSO variations.
Spectral analyses of the precipitation data show significant 11-year and 30-year periodicities in spring and summer respectively and some 3 to 4-year periodicity in the fall, winter, and spring . All the spectral properties of the Mongolian tree-ring series warrant further investigation and interpretation beyond the scope of this paper.