Canopy interception is a significant proportion of evapotranspiration of forest ecosystems

its quantification remains as one of hot issues in eco-hydrological studies. However

most studies were carried out just in the small scale of stand/plot. Since the canopy interception is influenced by many factors (e.g. rainfall

meteorological condition

canopy structure and growth process) at the stand/plot scale

and by the difference of forest canopy structure among slope positions

it often appears great spatial variation and scale effect at the slope scale. In order to realize the up-scaling from the measured value at plot to the whole slope

and to provide scientific basis for the accurate evaluation of forest hydrological impacts

the spatial-temporal variation and scale effect of canopy interception along the slope has to be precisely described based on deep-going studies. In this study

a representative slope with a horizontal length of 425.1 m and covered by Larix principis-rupprechtii plantation was selected in the small watershed of Xiangshuihe of Liupan Mountains in northwest China

and it was evenly divided into 16 continuous plots. In the growing season of 2015 (from May to October)

the dynamics of stand/plot canopy structure (especially the leaf area index

LAI) in different months and its difference among the 16 plots along slope position were monitored; meanwhile

the gross rainfall at open field and the throughfall and stemflow was measured to calculate the canopy interception of each plot during each rainfall event. The variation of canopy interception along slope positions and within different months and their main influencing factors were analyzed using the “horizontal distance from the slope top (horizontal slope length)” as a scale scalar. The results showed that the canopy interception presents a remarkable variation along the slope and spatial scale effect

with a remarkable difference among the months. During the whole growing season

the overall variation tendency of canopy interception with rising horizontal slope length showed firstly an increase

reaching its maximum at the middle slope

and then a decline. The scale effect was +5.62 mm/100 m within the horizontal slope length of 0~316.6 m and -2.37 mm/100 m within the horizontal slope length of 316.6~425.1 m. In May

the variation of canopy interception along rising horizontal slope length showed a decreasing tendency

with the scale effect of -0.26 mm/100 m within the horizontal slope length of 0~425.1 m. In June

July and August

the variation tendency of canopy interception along slope length was consistent with the whole growing season. In June

the scale effect was +1.28 and -1.78 mm/100 m within the horizontal slope length of 0~261.1 m and 261.1~425.1 m. In July

the scale effect was +0.92 and -0.88 mm/100 m within the horizontal slope length of 0~267.6 m and 267.6~425.1 m. In August

the scale effect was +1.28 and -0.34 mm/100 m within the horizontal slope length of 0~211.2 m and 211.2~425.1 m. In September and October

the variation of canopy interception along the slope showed an increasing tendency. The scale effect was +2.38 and +0.81 mm/100 m within the horizontal slope length of 0~425.1 m

respectively. The LAI was the main factor affecting the spatial-temporal variations of canopy interception. The correlation between canopy interception ratio and LAI was positive and significant in every month. The slope average of canopy interception could be estimated through the up-scaling from the measured LAI at certain plot

the relation between plot LAI at different slope positions and the slope average of LAI

and the relation between canopy interception and plot LAI.