Seasonal resource selection of free-ranging Zhongwei goats in the semi-arid grassland.

Goats rarely move and forage randomly. They tend to move in ways generally influenced by biotic and abiotic factors, respectively. However, few studies have explored the foraging behaviour of goats in the absence of predation and human disturbance. Based on step selection function modelling framework, Normalised Difference Vegetation Index, vegetation surveys, and Global Positioning System tracking of 124 free-ranging domestic adult male Zhongwei goats over one year (2016-2017) were used to assess how biotic and abiotic environmental factors affected their spatiotemporal distribution, and developed a conceptual model to represent the goats' trade-off between forage quantity and preference at different seasons, in the semi-arid grassland of Loess Plateau of 1 178 hectare. The results showed that spatial distributions of goats responded to spatiotemporal variation of biotic factors rather than abiotic factors of elevation, slope and solar radiation, which indicated that biotic factors were of priority to abiotic factors in the foraging process for the goats. According to the season changing, the goats positively used areas with higher forage quantity in the spring and winter, areas of higher forage quantity and preferred species in summer, and areas of abundance of preferred species in autumn. We developed a model to describe the phenomenon that the goats selected areas with higher preferred species only when the forage quantity was plentiful, otherwise they selected areas with higher forage quantity. Better understanding of the patterns and drivers of spatiotemporal distribution of the goats can improve our ability to predict foraging behaviour of livestock in heterogeneous environment and lead to better management practices and policies for the sustainability of these semi-arid landscapes and associated ecosystem services.

Legacy effects of precipitation amount and frequency on the aboveground plant biomass of a semi-arid grassland.

Precipitation is known to have legacy effects on plant diversity and production of many terrestrial ecosystems. Precipitation regimes are expected to become more variable with increasing extreme precipitation events. However, how previous-year precipitation regimes affect the current-year aboveground biomass (AGB) remains largely unknown. Here we measured long-term (2004-2017) AGB in a semi-arid grassland of the Chinese Loess Plateau to evaluate the impact of previous-year precipitation amount on current-year AGB. Furthermore, to assess the response of current-year AGB to previous-year precipitation regimes, we conducted a field manipulation experiment that included three precipitation regimes during 2014-2017: (i) ambient precipitation, (ii) monthly added four 5 mm rain events, and (iii) monthly added one 20 mm event. Both the long-term (2004-2017) observations under ambient precipitation and short-term (2014-2017) measurements under manipulative treatments showed significant positive effects of previous-year precipitation on current-year AGB. Our path analysis suggested that previous-year precipitation frequency had negative effects on the current-year density and mean height of grass (Leymus secalinus) while had positive effects on forb (Artemisia capillaris). The forb had much smaller height and AGB (65% and 53% less, respectively) than the grass. Consequently, the AGB reduced in the weekly small events treatment, causing the sensitivity of AGB to precipitation to decrease. Therefore, our findings indicated that the impacts of precipitation regimes on plant community dynamics should be taken into consideration while assessing the precipitation legacy effect on ecosystem production.

Impacts of climate change and human activities on grassland vegetation variation in the Chinese Loess Plateau.

China initiated the "Grain for Green Project" in 1999 to mitigate soil erosion. The vegetation cover of the Chinese Loess Plateau, one of the most erosive regions in the world, has been greatly increased. However, studies on quantitatively investigating the climate change and human activities on vegetation coverage change were rare. In this study, spatio-temporal changes in vegetation coverage were investigated using MODIS normalized difference vegetation index (NDVI) data over 2000-2016. And a new method was introduced using Net Primary Productivity (NPP) model and relationship between NPP and NDVI to quantitatively and spatially distinguish the NDVI affected by climate change and human activities. Results showed that mean NDVI value over 2009-2016 were 14.46% greater than that over 2000-2007. In order to quantify the contribution of climate change and human activities to vegetation change, an NPP model suitable for the grassland of the Chinese Loess Plateau was identified using biomass observations from field survey and literature. The NDVI affected by climate change (NDVIclimate) was estimated by the NPP model and the relationship between NPP and NDVI. And the NDVI affected by human activities (NDVIhuman) was calculated by actual NDVI minus NDVIclimate. Comparison of the two stages showed that human activities and climate change contributed 42.35% and 57.65% respectively to the ΔNDVI on grassland in the Loess Plateau. After analysis of numerous NDVIhuman related factors, the slopes restored by the "Grain for Green Project" was considered the main influence factor of human activities.

Under which climate and soil conditions the plant productivity-precipitation relationship is linear or nonlinear?

Understanding under which climate and soil conditions the plant productivity-precipitation relationship is linear or nonlinear is useful for accurately predicting the response of ecosystem function to global environmental change. Using long-term (2000-2016) net primary productivity (NPP)-precipitation datasets derived from satellite observations, we identify >5600pixels in the North Hemisphere landmass that fit either linear or nonlinear temporal NPP-precipitation relationships. Differences in climate (precipitation, radiation, ratio of actual to potential evapotranspiration, temperature) and soil factors (nitrogen, phosphorous, organic carbon, field capacity) between the linear and nonlinear types are evaluated. Our analysis shows that both linear and nonlinear types exhibit similar interannual precipitation variabilities and occurrences of extreme precipitation. Permutational multivariate analysis of variance suggests that linear and nonlinear types differ significantly regarding to radiation, ratio of actual to potential evapotranspiration, and soil factors. The nonlinear type possesses lower radiation and/or less soil nutrients than the linear type, thereby suggesting that nonlinear type features higher degree of limitation from resources other than precipitation. This study suggests several factors limiting the responses of plant productivity to changes in precipitation, thus causing nonlinear NPP-precipitation pattern. Precipitation manipulation and modeling experiments should combine with changes in other climate and soil factors to better predict the response of plant productivity under future climate.