Climate and soil texture controls of grassland ecosystem properties in Inner Mongolia from patch to regional scales

Long-term field monitoring, Century ecosystem model, remote sensing and geographically-referenced data are integrated to investigate quantitatively climate and soil texture controls on grassland ecosystem properties of Inner Mongolia (about 792,000 km{dollar}sp2{dollar} grasslands) from patch to regional scales. The overall objective is to quantify spatial-temporal climate variability and to determine the relative importance of monthly and seasonal precipitation and temperature in controlling dynamics of plant biomass, primary production, water use efficiency and soil organic matter dynamics at patch scale and spatial distribution of vegetation at regional scale.; On the regional scale, annual precipitation of 156 weather stations ranged from 37.9 mm in the west to 538.5 mm in the northeast over the period of 1950-1980. Precipitation in April-September accounted for 84% to 92% of annual precipitation. The coefficient of variation (CV) in annual precipitation in 1971-1980 varied from 15% in areas of high annual precipitation to 40% in areas of low annual precipitation. There are better geographical correspondences between vegetation patterns and patterns of monthly precipitation and monthly minimum temperature, compared to annual precipitation and mean annual minimum temperature. Precipitation in April-September and monthly minimum temperature in January were better indicators of vegetation distribution than annual precipitation and mean annual minimum temperature.; At the patch scale, the field monitoring during 1980-1989 in the Xilin river basin showed that the CV in peak aboveground live biomass (PALB) and peak standing crop (PSC) were 29% and 26% for Leymus chinense steppe, 24% and 25% for Stipa grandis steppe, while the CV in annual precipitation was 22%. On the average, PALB and PSC of L. chinense steppe were 182.68 g.m{dollar}sp{lcub}-2{rcub}{dollar} and 193.48 g.m{dollar}sp{lcub}-2{rcub}{dollar}, while those of S. grandis steppe were 144.43 g.m{dollar}sp{lcub}-2{rcub}{dollar} and 152.12 g.m{dollar}sp{lcub}-2{rcub}{dollar}, probably due to soil texture difference. Using PSC as estimate of annual aboveground net primary production, rain use efficiency of L. chinense steppe was 6.3 kgDM.ha{dollar}sp{lcub}-1{rcub}{dollar}mm{dollar}sp{lcub}-1{rcub}{dollar}yr{dollar}sp{lcub}-2{rcub}{dollar} and S. grandis steppe 4.9 kgDM.ha{dollar}sp{lcub}-1{rcub}{dollar}mm{dollar}sp{lcub}-1{rcub}{dollar}yr{dollar}sp{lcub}-1{rcub}{dollar}. Aboveground biomass of various species of L. chinense steppe had different responses to interannual variations of precipitation and temperature during 1980-1989. Century model simulations were slightly more successful than the statistical models using annual precipitation in estimating annual primary production of L. chinense steppe and S. grandis steppe. Using Landsat TM images on 31 July 1987 and 11 August 1991, NDVI-derived aboveground standing biomass were within {dollar}pm{dollar}25% of the observed data.