Loess on the high mountains of northwestern China is largely controlled by the topographic features, distributed mainly on the northern slopes of the Tianshan and Kunlun Mountains as well as in the Ili Valley (between the northern and central Tianshan ranges). Loess distributed in different regions has different ages, with the oldest (early Pleistocene) loess occurring in the Ili Valley. Geochemical, mineralogical and granulometric analyses demonstrate that the loess in the three different regions has different provenance. The loess on the northern Tianshan and Kunlun Mountains is mainly derived from the Junggar and Taklimakan Deserts, respectively, whereas the loess in the Ili Valley is derived mainly from the Sary–Ishikotrau Desert located in the Republic of Kazakhstan. However, these deserts serve mostly as holding areas for the silts rather than original source regions. The production of the silt-sized particles is mainly associated with glacial abrasion and tectonic-induced rock denudation. Thus, the loess in the studied regions can be described as “mountainous” loess rather than simply “glacial” loess or “desert” loess. It is unlikely that the mechanisms of loess formation are the same in different regions, given their dependence on specific geomorphological conditions and associated processes.

Plant functional types across environmental gradients can be considered as a powerful proxy that reveals vegetation–environment relationships. The objectives of this study were to investigate the response in the relative abundance of plant functional types along altitudinal gradients and to examine the relationship of plant functional types to environmental variables. The study was conducted in the Gughe-Amaro Mountains, in the south-west Ethiopian highlands. We established 74 plots with an area of 400 m2 (20 × 20 m) each along altitudinal ranges between 1000 and 3000 m asl. Data on site environmental conditions and on the abundance of plant functional types were analysed using the constrained linear ordination technique (RDA) in order to identify the relationships between plant functional types and environmental variables. Altitude, soil organic carbon, soil sand fraction and surface stone cover were significantly related to the relative abundance of plant functional types across the gradient. Tussocks and thorns/spines were abundant in lower altitudinal ranges in response to herbivory and drought while rhizomes and rosettes were abundant at higher altitudes in response to the cold. Generally our results show that topographic attributes (altitude and slope) as well as soil organic carbon play an important role in differentiating the relative abundance of plant functional types in the investigated gradient. Thus, considering specific plant functional types would provide a better understanding of the ecological patterns of vegetation and their response to environmental gradients in tropical regions of Africa prone to drought.