Topography-Controlled Soil Water Content and the Coexistence of Forest and Steppe in Northern China

1

Physical Geography, 2012, 33, 5, pp. 1–XX. http://dx.doi.org/10.2747/0272-3646.33.5.1Copyright © 2012 by Bellwether Publishing, Ltd. All rights reserved.

TOPOGRAPHY-CONTROLLED SOIL WATER CONTENT AND THE COEXISTENCE OF FOREST AND STEPPE IN NORTHERN CHINA

Hongyan Liu1

Laboratory for Earth Surface Processes, Ministry of Education, and College of Urban and Environmental Sciences

Peking University, Beijing, 100871 China

Siyuan HeLaboratory for Earth Surface Processes

Peking University Beijing, 100871 China

Oleg A. AnenkhonovInstitute of General and Experimental Biology Russian Academy of Sciences, Siberian Branch

Ulan-Ude 670047, Russia

Guozheng HuLaboratory for Earth Surface Processes

Peking University Beijing, 100871 China

Denis V. Sandanov and Nathalia K. BadmaevaInstitute of General and Experimental Biology Russian Academy of Sciences, Siberian Branch

Ulan-Ude 670047, Russia

Abstract: The semi-arid forest-steppe ecotone in China is characterized by a patchy pat-tern of forest and steppe, with forest patches restricted to shady slopes. To address the effect of topography on forest distribution through regulation of available water, we calculated evaporation as a function of slope aspect and inclination. Field vegetation records from randomly selected sites with minimum slope inclination were used to test the simulated forest distribution. Seasonal and diurnal changes of surface soil temperature and mois-ture of shady and sunny slopes were recorded. Soil water content was measured during two growing seasons on both sunny and shady slopes with the same forest type at three sites located along the mean annual precipitation (MAP) gradient. Evaporation decreases with slope inclination on shady slopes, but increases with inclination on sunny slopes. The shady slope received 35% of the annual direct solar radiation received by the sunny slope when the slope inclination was 25°, and the contrast in annual direct solar radiation between the shady and sunny slopes further widens as slope inclination increases. Steeper shady slopes can support forests in dryer climates, with log-linear regression revealing a

1Corresponding author; email: [email protected]

[Authors: Please check that author affiliations have been indicated correctly in the new format]

2 LIU ET AL.

minimum slope inclination for forest distribution along the MAP gradient. The simulated minimum slope inclination for forest growth was larger than the observed minimum incli-nation, and the difference was greater in wetter conditions. A larger forest area fraction was considered to lead to a reduction in soil temperature and evaporation, as verified by soil temperature and moisture records and soil water content measurements. The slope-specific forest distribution in the semi-arid region of China can be explained by a topography- controlled soil water supply. Lower evaporation, resulting from lower direct solar radiation on shady slopes, allows shady slopes to retain a water supply sufficient for sustaining for-ests, and the existence of forests on shady slopes further reduces evaporation. Different tree species coexist at the xeric timberline due to regulation by slope inclination and aspect. [Key words: forest-steppe, ecotone, topography, East Asian steppe, soil moisture, slope aspect.]

INTRODUCTION

Moisture-limited vegetation in semi-arid regions supports approximately 36% of the human population and maintains considerable biodiversity (Bailey, 2010). For-ests in semi-arid regions contribute significantly to the global carbon budget through their higher than average carbon storage efficiency (Rotenberg and Yakir, 2010). In the coming decades, global climate change is expected to result in large shifts in vegetation distributions at unprecedented rates, particularly in semi-arid landscapes (Allen and Breshear, 1998; Frelich and Reich, 2010; Hirota et al., 2010). Dynamic global vegetation models (DGVMs) also predict that regional changes in precipita-tion may lead to a net decline of woody vegetation cover in many semi-arid regions (Lucht et al., 2006). Previous predictions by different models have so far merely focused on the shifting of forest vegetation, instead of considering changes to vegeta-tion patterns in the semi-arid regions, mainly due to a lack of field data. Such short-comings limit the application of these models in the forest management of semi-arid regions. Regional-scale data from field studies are thus urgently needed to provide a basis for predicting the response of semi-arid forests to climate changes (Allen and Breshear, 1998).

The temperate forest-steppe ecotone is a transition between broadleaved decidu-ous forest and temperate steppe, and is at the “drought limit” of tree survival as well as forest distribution (Bugmann, 1996). The dominant forest-steppe landscape in Inner Asia, including southern Siberia (Russia), northern Mongolia, and Inner Mongolia (China) is characterized by treeless south-facing slopes and forested north-facing slopes (Liu et al., 2000; Dulamsuren et al., 2005a, 2005b; Namzalov, 2009), differing from the tropical savannas (Sankaran et al., 2005). A debate con-tinues on the formation of this unique landscape, which has been proposed to be accounted for either by large-scale human deforestation (Hilbig, 1995, 2003), or by climate (Gunin et al., 1999; Dulamsuren et al., 2005a, 2005b). Recent work in the boreal forest-steppe ecotone has tended to support climate-driven, aspect-based forest-steppe patterns in southern Siberia and northern Mongolia, but has not denied that in populated areas grazing may have increased the proportion of steppe at the expense of forest (Chytrý et al., 2008; Schlütz et al., 2008). Although topography and aspect have been studied in other ecotones (e.g., Goldblum et al., 2010), so far no

FOREST-STEPPE ECOTONE 3

quantitative investigations show how topography regulates the aspect-based distri-bution of forests within the forest-steppe ecotone.

It is generally regarded that tree survival and growth at the xeric limit of trees is susceptible to the regional water balance (Bailey, 2000). Woody cover in the African savanna was demonstrated to be determined by mean annual precipitation (MAP) regimes (Sankaran et al., 2005). We therefore propose that MAP might also determine woody cover in the forest-steppe ecotone. However, the forest in the temperate forest-steppe ecotone is quite different from that of the tropical savanna in terms of its vegetation structure and distribution. Solar radiation, the most evident habitat difference between shady and sunny slopes, might lead to an aspect-based forest distribution controlled by differentiated light conditions or soil water contents caused by differentiated evaporation between shady and sunny slopes. If the former was the case, then shade tolerant rather than light-preferring tree species would be expected to survive on shady slopes, and there will be no difference in tree species along the MAP gradient. Based on a common recognition of the water susceptibil-ity of forests at their xeric timberline, we hypothesize that topography regulates soil water supply and further determines forest distribution within the forest-steppe ecotone.

In this paper, we report results from a model developed to calculate evaporation changes and predict forest distribution in terms of slope inclination and aspect. Field records of forest distribution and soil moisture variability under different topographi-cal situation were employed to test the model simulation. We explain how topog-raphy controls soil water supply and, subsequently, the vegetation patterns, using a case study in southeastern Inner Mongolia, China.

STUDY AREA

The study area was located in the region 116º20′ E to 117º50′ E and 41º55′ N to 43º35′ N. The region is situated mostly on the southeastern Inner Mongolian Plateau with elevations ranging from 1100 m to 1400 m a.s.l. Landforms on the Plateau sur-face were varied. The Otindag (Hunshandake) Sandy Land, consisting of vegetation-covered sand belts, is located in the east, whereas the Xilinguole Lava Platforms are located in the north. Lakes and hills are scattered throughout the plateau. Mountains with elevations of up to 2000 m a.s.l. border the Inner Mongolian Plateau on both the east and south (Fig. 1).

The study area lies in a transitional zone from a temperate semi-humid monsoon climate to a semi-arid continental climate. Soil and vegetation change along the climatic gradient. From the Jibei Mountains in the southeast to the Xilinguole Lava Platform in the northwest, soil types change in the order of brown soil, grey forest soil, chernozem, light chernozem, and dark kastanozem, according to the soil clas-sification system widely adopted in China (Xiong and Li, 1987). Otindag Sandy Land consists of vegetation-covered dunes and has an aeolian sandy soil. Vegetation types change from forest in the southeast to steppe in the northwest, with either a broader or narrower ecotone between them.

4 LIU ET AL.

METHODS

Vegetation Surveys

To quantify the linkage between vegetation distribution and topography, we sys-tematically selected 77 sites in the forest-steppe ecotone, encompassing different forest types and topographical conditions (Fig. 1). We randomly selected 149 plots for both forests (10 × 10 m) and steppes (4 × 4 m) from different topographic slopes available locally. We recorded species composition and topographic features in each plot. We classified forest types according to dominant tree species.

Satellite Image Interpretation

We interpreted forest fractional cover along a precipitation gradient from satellite images. ETM+ data from Landsat-7 (July 1, 1999 and July 6, 2001, track numbers

Fig.1. Location of the study area. Circles denote sites with vegetation plots and crosses denote sites with soil samples. Precipitation patterns in the study area and its surroundings, interpreted from 31 mete-orological stations by kriging, are shown by isohyets (dashed lines). The codes A, B, and C indicate the three sites chosen for soil water content measurements.


123–30 and 123–31, geometrically corrected) were joined into one image. The two images had little cloud cover, and vegetation was in its full growth stage, reflect-ing the true vegetation cover. We manually identified forest patches in the satellite image in its false-color mode, as forests are distinctly different from grasslands in this mode. The quality of this delineation was tested and improved through field observa-tion at different parts of the study region. We measured forest fractional covers for different mean annual precipitation (MAP) ranges at 10 mm intervals.

Interpolation of Climate Variables

We collated annual precipitation and mean annual temperature (MAT) for the period 1980 to 2000 at 31 meteorological stations in the study area and the sur-rounding territory to establish the relationship between climatic parameters (MAP and MAT) and geographical parameters (longitude, latitude, and altitude).

Solar radiation is the strongest driving force of evaporation, so we extracted evaporation and solar radiation data from the eight meteorological stations along a latitudinal gradient in China, from the year 1951 to the year 2003. The following regression equation was calculated:

E = 152.02e0.0143S′ R2 = 0.607 P = 0.023 (1)

where E is annual evaporation and S′ is direct solar radiation, which is a function of slope aspect and inclination.

We adopted the optimal equation (2) following Weng (1986), who compared several equations for calculating direct solar radiation in China, and found:

S′ = S0 (as1 + bs12) (2)

where S′ is direct solar radiation as in equation (1), S0 is astronomical solar radiation, a = 0.281 and b = 0.334 are empirical coefficients, and s1 is percentage sunshine duration. The average relative error of this formula is 5.8% (Weng, 1986).

We calculated astronomical solar radiation in the study area from the distribution of astronomical solar radiation in the range 0°N–90°N, using:

S0 = 0.0009Lat5 – 0.1272Lat2 + 1.4484Lat + 421.19 R2 = 0.9988 (3)

where S0 is astronomical solar radiation and Lat is latitude. In our study region, s1 is 71% (http://www.naturalresources.csdb.cn/zrzy/ntBC02.asp?Page=4). Therefore S0 is 323.69 W/m2 according to (3) and S′ is 120.71 W/m2 according to (2). Then we corrected S′ for different slope aspects and inclination using the ARCGIS com-mand HILLSHADE, and calculated annual evaporation for different slope aspects and inclinations using equation (1).

Measurement of Soil Temperature and Moisture

We installed two U23-001 HOBO Pro v2 data loggers under the soil surface at a depth of 10 cm in the forest and steppe sites, respectively, in Baiqiankeng (Site 1 in Fig. 1). Soil temperature and moisture were measured every half hour. We calculated

6 LIU ET AL.

daily temperature and moisture values to show variations between the forest and neighboring steppe. The measured moisture is expressed as humidity of the air in the soil pore space rather than soil water content.

We collected soil samples from three sites with Populus davidiana forest on the shady slopes and steppe on the sunny slopes in July, August, and September of 2007, and May and June of 2008 (Fig. 1). In each plot, we excavated two soil profiles in both the shady and sunny slopes and collected soil samples along the slope profile at an interval of 20 cm from 10 to 70 cm under the surface, each sample filling a 100 cm3 cutting ring. We measured water contents in the lab. Each soil sample was weighed, put into aluminum boxes, and dried at 105°C in the oven for 5 hours to constant weight. We then calculated the soil water content.

Statistical Analysis of Forest Distribution with Topography

Field records were used to investigate the relationship between slope inclination and MAP for forested sites. Quantile regression (Cade and Noon, 2003) was used to determine the minimum and average slope inclinations for tree survival at different MAPs. The regression results were compared with simulated water conditions under different MAPs and slope inclinations.

RESULTS

Distribution Patterns of Precipitation and Simulated Evaporation

Temperatures were relatively higher in the valley and lower on the high peaks of the mountainous areas; however, there was no marked variation across the Plateau (results are therefore not plotted here). MAP exhibited a NW-SE gradient (Fig. 1). From the mountains at the edge of the Inner Mongolian Plateau in the southeast to the Xilinguole Lava Platform in the inner region of the Plateau, MAP decreased from 450 mm to 320 mm. MAP showed a significant correlation with longitude, latitude, and altitude (p < 0.01):

P = –878.918 – 56.657Lat + 31.676Long – 0.008Alt n = 31, R2 = 0.74 (4)

An opposite trend was found between evaporation and slope inclination. As slope inclination increases, the difference in evaporation between shady and sunny slopes widens (Fig. 2). The evaporation from north facing slopes is 80.1% of that from south facing slopes when the slope is 5º, but decreases to 22.4% when the slope inclina-tion increases to 40º.

Observed Relations between Vegetation Distribution and Topography

The recorded 55 forest plots are constrained by slope aspect, with 30 directly fac-ing north, 15 facing NE, and eight facing NW. Only one plot faces east and one faces west; these have close to the highest MAP. Steppe distribution, however, includes all inclinations.

Quantile regressions of MAP and slope inclination with forest distribution are shown in Figure 3. At sites where MAP exceeds 450 mm, the forest is distributed


over a range of slope inclinations. Within the MAP range of 340 to 450 mm, slopes supporting forest become steeper with decreasing MAP, as indicated by the signifi-cant regression between MAP and minimum slope inclination for the 0.9 quantile of forest distribution (y = –161.01log(x) + 433.2, p < 0.05). The average slope indi-cated by regression with the 0.5 quantile of forest distribution is y = –122.45log(x) + 345.35 (p < 0.01).

When we compared these results with the modeled soil water content at roughly equal MAP and evaporation values (y = -50.37log(x) + 332.8; p < 0.001), we found that sites with the average slope inclination required for forest distribution were quite close to sites with a net forest moisture budget of zero. The minimum observed slope inclination required for forest distribution, however, is lower than that of the simu-lated slope inclination, and the discrepancy between the model and observations increases as MAP increases (Fig. 3).

No clear link between MAP and dominant tree species distribution was found. Six forest types, dominated by Betula platyphylla, Betula davurica, Populus davidiana, Quercus mongolica, Pinus tablaeformis, and Picea meyeri, were identified within the ecotone. Although three of these are closer to the xeric timberline in the driest forested sites, the other three species were also found as non-dominant species at these sites (Fig. 3). A reduction in average forest fractional cover with decreasing precipitation was found (p < 0.01; Fig. 4).

Variation in Soil Temperature and Moisture with Topography

Measured soil water contents show that the soil water contents at different depths decrease as MAP decreases. The differences in soil water content between forest on shady slopes and steppe on sunny slopes become smaller at drier sites (Fig. 5A).

Fig. 2. Annual evaporation patterns for different slope exposures. The ratio of shady (north) slope evaporation to sunny (south) slope evaporation is shown as a percentage indicated on the right axis.

8 LIU ET AL.

The growing season (April to October) soil temperature is higher in the steppe than in the forest, and vice versa in the non-growing season. The highest temperature difference reaches about 12º. The recorded soil moisture is higher in the forest than

Fig. 3. Inclination of forested slopes along the MAP gradient. Different forest types are indicated by circles (Betula davidiana), diamonds (Betula platyphylla), crosses (Populus davidiana), squares (Picea meyeri), triangles (Pinus tabulaeformis), and hexagons (Quercus mongolica). The 0.9 percentile regression between minimum slope inclination for tree survival and MAP follows the equation y = –161.01log(x) + 433.2 (p < 0.05), as indicated by a dark solid line; the 0.5 percentile regression between average slope inclination and MAP is y = –122.45log(x) + 345.35 (p < 0.01), as indicated by a light solid line. Slope inclination predicted for when annual evaporation equals MAP is indicated by “+” symbols, and the dashed line shows regression between the predicted slope inclination and MAP (y = –50.37log(x) + 332.8; p < 0.001).

Fig. 4. Change of forest fractional cover as interpreted from satellite images, along the MAP gradient. Each point represents a forest patch area, with precipitation increments of 10 mm in the range from 340 to 460 mm.


in the steppe throughout the year except at the beginning and end of the growing season (Fig. 5B).

DISCUSSION

Our study indicated reduction in both forest fractional cover and patch size with decreasing precipitation. The vegetation gradient differed markedly from that of tropical savannas where grasses and woody plants occupy the same space, and woody plant coverage varies along the precipitation gradient (Sankaran et al., 2005).

Fig. 5. A. Soil water content variations at different depths at three sites dominated by Populus davidi-ana along the MAP gradient. Measured soil water contents of two profiles at each forest site (indicated as N) and the neighboring steppe site (indicated as S) for different months during two growing seasons are averaged. Locations of sites A–C are shown in Figure 1. For each site, bars from left to right represent soil water content at depths of 10 cm, 30 cm, 50 cm, and 70 cm below the surface. B. Difference in top soil temperature and moisture between steppe and forest as recorded by a U23-001 HOBO Pro v2 data logger in the year 2009 to 2010. Only the daily mean values of soil temperature and moisture for the 5th, 15th, and 25th days of each month are presented. Measurements started on November 5, 2008 and ended on November 25, 2009.

10 LIU ET AL.

In contrast, the landscape was characterized by closed-canopy forest patches sur-rounded by a matrix of steppe vegetation, and forest fractional cover decreases along the precipitation gradient (Figs. 1 and 4). Even at the driest sites, trees formed a closed canopy (Liu et al., 2000). The decreasing forest patch size along the precipita-tion gradient implied that forests might respond to climate drying by reducing their patch size, an adjustment that has not been considered in previous prediction on dynamics of semi-arid forests (Frelich and Reich, 2005).

The link between forest cover gradient and bioclimatic gradient in our study high-lights the effect of drought stress, mainly driven by soil water content. Despite the differences in vegetation patterns, the principal determinant of the vegetation pattern in the temperate forest-steppe ecotone in our study region appears to be available water, which was also the primary factor explaining change in woody cover in African savannas (Sankaran et al., 2005). A remarkable characteristic in our study area is that MAT varied very little, whereas MAP varied markedly (Fig. 1). Meanwhile, sunny slopes received more solar radiation, which led to greater evaporation, whereas reduction in effective solar radiation led to a marked decrease of evaporation on shady slopes (Fig. 2). These topographic effects seem to be the key factor leading to greater available soil water on steep shady slopes, which, in turn, promotes the growth of detectable forest vegetation. As our study area is relatively uniform in topography and there is no relationship between MAP and topographic slope (e.g., high slope areas occur only at low MAP), we can confirm the relationship between slope, evaporation, and soil available water for forest survival. Our results provide further evidence for a climate-determined, aspect-based forest-steppe landscape in Inner Asia (Gunin et al., 1999; Dulamsuren et al., 2005a, 2005b).

The effect of low light conditions on the shady slope did not necessarily favor for-est growth vis-à-vis steppe. Light-preferring tree species, such as birch and poplar, are all widely distributed on shady slopes, even close to the xeric timberline (Fig. 4). A minimum slope inclination for forest survival due to more water consumption by forests than by steppes is thus suggested. It is also possible that higher radiation in sunny areas favors the steppe species relative to any of the forest species (likely at the seedling stage), leading to the competitive dominance of steppe vegetation on sunny slopes (Gómez-Aparicio et al., 2004). However, this effect did not contradict the limitation of water on adult tree survival on sunny slopes.

The effect of topography on soil water content, however, is also greatly affected by the features of vegetation. The minimum slope inclination required for tree survival is much lower than that simulated using the net soil water content at the higher end of the MAP range within the forest-steppe ecotone. This finding may be because the forest itself can reduce evaporation, due to decreased soil temperature under the forest canopy. A higher fraction of forest cover in areas with greater MAP may regulate water loss and permit forest to thrive on gentle shady slopes, a hypothesis supported by the higher contrast in soil water content between forest and steppe at wetter sites with a higher forest fraction (Figs. 4 and 5A). Further evidence for the contribution of forest fraction to evaporation reduction is in the different seasonal patterns in soil temperature and moisture between sunny and shady slopes. Both soil temperature and soil moisture were lower in the steppe than in the forest during the non-growing season (Fig. 5B). Soil temperature was higher in the steppe than in


forest, whereas soil moisture showed the reverse pattern during the growing season (Fig. 5B). Although we observed soil temperature and moisture in only one repre-sentative site in our study area, unpublished data obtained in our wide observation from southeastern Inner Mongolia to southern Siberia (42o–55o N) shows a repeated pattern. In our study, we only calculated potential evaporation under different slopes and gradients. Soil water content, however, is determined also by transpiration of the vegetation itself. As forests generally have greater water demand than grasslands (Jackson et al., 2005), the high water content of forest soil during the growing season further indicated the role of topography in regulating evaporation.

Different soil properties under shady and sunny slopes might also contribute to soil water content. Systematic sampling in the semi-arid region in Boise, Idaho, USA showed that soils on the shady slope retain as much as 25% more water at any given soil water pressure than samples from the sunny slope, which is accounted for by greater soil thickness, soil porosity, and soil organic matter and silt content on the shady slope than on the sunny slope (Geroy et al., 2011). Although we have not sys-tematically investigated soil physical and chemical properties in this study, we can infer coarser soil texture and lower soil thickness on sunny slopes, owing to stronger wind erosion and lower vegetation cover on sunny slopes than on shady slopes. Gravitational drainage is a very important factor in regulating soil water content (Geroy et al., 2011). Due to this effect, forests are mostly distributed on middle to lower slopes than on upper slopes. Because we focused on the difference between shady and sunny slopes, we did not calculate the topographic moisture index under different slope positions.

Other abiotic factors such as snow and wind might also contribute to the distribu-tion of forest within the forest-steppe ecotone (Coupland, 1979). However, these fac-tors may also be controlled by topographic conditions to some extent (cf. Holtmeier and Broll, 2010, for treeline ecotones). As we observed in the field, snow depth was lower on sunny slopes than on shady slopes due to faster melting as well as increased transportation by wind. Forests on shady slopes, on the other hand, may prevent snow from melting and prevent aeolian transportation of snow, which, in turn, contributes to greater soil moisture during the growing season.

Our simulation and observations appear to support our hypothesis that topogra-phy-controlled soil water content determines the coexistence of forest and steppe in southeastern Inner Mongolia, China. Previous study suggested two types of veg-etation response to future climate change in the forest-prairie ecotone in North America—either savannification with trees thinning within forest patches or retreat of forest distribution (Frelich and Reich, 2005). Our study suggests another alternative of forest-steppe ecotone response (e.g., reduction in slope-scale forest patch sizes) to climate change. The vegetation zone distributions might remain stable under a changing climate, due to the buffering effect of topography. To make a more precise prediction of the future dynamics of a forest-steppe landscape that is susceptible to climate change, we need to incorporate topography-controlled soil water content in vegetation models.

Acknowledgments: This work is supported by a grant from the National Natural Science Foundation of China (NSFC) project (No. 41071124) and a joint project between the NSFC and the Russian Foundation

12 LIU ET AL.

for Basic Research (RFBR) (No. 40711120173). The authors also thank the China Meteorological Admin-istration (CMA) for providing data.

REFERENCES

Allen, C. D. and Breshear, D. D. (1998) Drought-induced shift of a forest–wood-land ecotone: Rapid landscape response to climate variation. Proceedings of the National Academy of Sciences, Vol. 95, 14,839–14,842.

Bailey, R. M. (2010) Spatial and temporal signatures of fragility and threshold prox-imity in modelled semi-arid vegetation. Proceedings of the Royal Society B: Biological Science, Vol. 278, 1064–1071.

Bugmann, H. K. M. (1996) Functional types of trees in temperate and boreal forests: Classification and testing. Journal of Vegetation Science, Vol. 7, 359–370.

Cade, B. S. and Noon, B. R. (2003) A gentle introduction to quantile regression for ecologists. Frontiers in Ecology and Environmental Sciences, Vol. 1, 412–420.

Chytrý, M., Danihelka, J., Kubešová, S., Lustyk, P., Ermakov, N., Hajek, M., Hajkova, P., Koci, M., Otypkova, Z., Rolecek, J., Reznickova, M., Smarda, P., Valachoic, M., Popov, D., and Pisut, I. (2008) Diversity of forest vegetation across a strong gradi-ent of climatic continentality: Western Sayan Mountains, southern Siberia. Plant Ecology, Vol. 196, 61–83.

Coupland, R. T. (1979) Grassland Ecosystem of the World: Analysis of Grasslands and Their Uses. New York, NY: Cambridge University Press.

Dulamsuren, C., Hauck, M., and Mühlenberg, M. (2005a) Ground vegetation in the Mongolian taiga forest-steppe ecotone does not offer evidence for the human origin of grasslands. Applied Vegetation Science, Vol. 8, 149–154.

Dulamsuren, C., Hauck, M., and Mühlenberg, M. (2005b) Vegetation at the taiga forest-steppe borderline in the western Khentey Mountains, northern Mongolia. Annales Botanici Fennici, Vol. 42, 411–426.

Frelich, L. E. and Reich, P. B. (2010) Will environmental changes reinforce the impact of global warming on the prairie–forest border of central North America? Frontiers in Ecology and the Environment, Vol. 8, 371–378.

Geroy, I. J., Gribb, M. M., Marshall, H. P., Chandler, D. G., Benner, S. G., and McNamara, J. P. (2011) Aspect influences on soil water retention and storage. Hydrological Processes, Vol. 25, 3836–3842.

Goldblum, D., Rigg, L. S., and Napoli, J. M. (2010) Environmental determinants of tree species distributions in central Ontario, Canada. Physical Geography, Vol. 31, 423–440.

Gómez-Aparicio, L., Zamora, R., Gómez, J. M., Hódar, J. A., and Castro, J. (2004) Applying plant facilitation to forest restoration: A meta-analysis of the use of shrubs as nurse plants. Ecological Application, Vol. 14, 1128–1138.

Gunin, P. D., Vostokova, E. A., Dorefeyuk, N. I., Tarasov, P. E., and Black, C. C. (1999) Vegetation Dynamics of Mongolia. Dordrecht, Netherlands: Kluwer.

Hilbig, W. (1995) The Vegetation of Mongolia. Amsterdam, Netherlands: SPB Aca-demic Publishing.


Hilbig, W. (2003) Forest distribution and retreat in the forest steppe ecotone of Mongolia. Marburger Geographischer Schrift, Vol. 135, 171–178.

Hirota, M., Nobre, C., Oyama, M. D., and Bustamante, M. M. C. (2010) The climatic sensitivity of the forest, savanna and forest–savanna transition in tropical South America. New Phytologist, Vol. 187, 707–719.

Holtmeier, F-K. and Broll, G. (2010) Wind as an ecological agent at treelines in North America, the Alps, and the European subarctic. Physical Geography, Vol. 31, 203–233.

Jackson, R. B., Jobbagy, E. G., Avissar, R., Roy, S. B., Barrett, D. J., Cook, C. W., Farley, K. A., le Maitre, D. C., McCarl, B. A., and Murray, B. (2005) Trading water for carbon with biological carbon sequestration. Science, Vol. 310, 1944–1947.

Liu, H., Cui, H., Pott, R. and Speier, M. (2000) Vegetation of the woodland-steppe ecotone at the southeastern edge of the Inner Mongolian Plateau. Journal of Vegetation Science, Vol. 11, 525–532.

Lucht, W., Schaphoff, S., Erbrecht, T., Heyder, U., and Cramer, W. (2006) Terrestrial vegetation redistribution and carbon balance under climate change. Carbon Bal-ance and Management, Vol. 1, doi:10.1186/1750-0680-1-6.

Namzalov, B. B. (2009). Baikal phytogeographic node as the newest center of ende-mism of Inner Asia. Contemporary Problems of Ecology, Vol. 2, 341–347.

Rotenberg, E. and Yakir, D. (2010) Contribution of semi-arid forests to the climate. Science, Vol. 327, 451–454.

Sankaran, M., Hanan, N. P., and Scholes, R. J. (2005) Determinants of woody cover in African savannas. Nature, Vol. 438, 846–849.

Schlütz, F., Dulamsuren, C., Wieckowska, M., Mühlenberg, M., and Hauck, M. (2008) Late Holocene vegetation history suggests natural origin of steppes in the northern Mongolian mountain taiga. Palaeogeography, Palaeoclimatology, Pal-aeoecology, Vol. 261, 203–217.

Weng, D. M. (1986) Climatic method for direct solar radiation calculation and its distribution over China. Acta Energiae Solaris Sinica, Vol. 7, 122–130 (in Chinese with English abstract).

Xiong, Y. and Li, Q. (1987) Soils of China. Beijing, China: Science Press (in Chinese).

Topography-Controlled Soil Water Content and the Coexistence of Forest and Steppe in Northern China

Documents