Fig 1.
Location map of the study area.
Fig 1 (a) and (b) show the location of research area in the maps of Asia and Guiyang Province, respectively. For Fig 1(a), yellow area shows location of Guizhou. For Fig 1(b), red boundary shows research watershed selected. Yellow area represents karst area and green area represents non-karst area. For Fig 1(c), hydrologic station (red triangle), rainfall station (blue polygons), study area (red line), county border (black dotted line), and river system (blue line) in the study area are presented. And Digital Elevation Model (DEM) also shows on the map. Maps in Fig 1 were generated by ArcGIS 10.2 using the free download data online of the Geospatial Data Cloud (http://www.gscloud.cn/) and URL link of the software is http://www.esri.com/.
Fig 2.
The technology roadmap of the whole article.
Fig 3.
The linear trend of annual temperature.
In the figure, green and yellow lines with markers respectively represent annual average temperature and the moving average of 5a of temperature. And green dotted line is the trendline of annual average temperature.
Fig 4.
Anomalous variation of temperature.
In the figure, red and gray bars respectively represent positive and negative departures from average value.
Fig 5.
Scatterplot of mean annual temperature and annual runoff.
Fig 6.
The interannual variations of runoff and precipitation in 46 years (a) and the annual variation curve of the average runoff and precipitation for several years (b).
In Fig 6(a), blue and red bars respectively represent runoff and precipitation. And blue dotted line, red dotted line are respectively the trendline of them. In Fig 6(b), blue bars and red line respectively show the annual distribution of runoff and precipitation, and only the percentage of monthly runoff are showed in the picture.
Fig 7.
Time-frequency structure figure of the real part of Morlet wavelet transform coefficients of annual precipitation (a) and runoff (b).
Fig 7 describe the high and low flow phase structures under different time scales by employing Matlab 7.0. And tone from warm to cold represent high values to low values of the real part of Morlet wavelet transform coefficients.
Fig 8.
Wavelet variance diagram of annual precipitation (a) and runoff (b).
In figures, black lines paint the fluctuation of wavelet variance with scale. Red numbers and arrows point out the peak of variance which corresponds to different primary periods of series from large to small.
Fig 9.
Trend of changes in annual precipitation (a) and runoff (b) on scales of major periods.
For Fig 9(a), blue, red and yellow lines respectively represent the main periods 3a, 9a and 23a. And for Fig 9(b), blue and red lines respectively represent the main periods 3a and 14a. The main periods were got from the verification result of wavelet variance of Fig 8.
Fig 10.
Normalized curve of annual runoff and annual precipitation.
This figure shows the multiyear variation curve after standardization treatment of annual runoff (yellow area) and annual precipitation (green line) in study area. Time periods 1968–1980, 1981–2006 and 2007–2013 are represented by period I, II and III. Period divided on the basis of the goodness of fit between annual runoff and annual precipitation. Annual runoff and precipitation exhibit a good synchronous change relationship in the fluctuation in period II and show a clear separation in period I and III.
Table 1.
Variations of annual streamflow and precipitation in Liudong River watershed.
Fig 11.
Spatial distribution maps of land use in 1990 (a), 2000 (b) and 2010 (c).
Forestry, grass land, water area, arid land, paddy field and building lot were classified in maps. Overall accuracy values of three-phase land use vector diagram are 96.89%, 98.55% and 93.17%, and kappa coefficient values are 0.95, 0.97 and 0.93. Maps in Fig 11 were generated by ArcGIS 10.2 and ENVI 5.2 and the URL links of software are http://www.esri.com/ of both. Three-phase TM images of 1990, 2000 and 2010 from free download data of U.S. Geological Survey (http://glovis.usgs.gov/) were used.
Table 2.
Area and proportion of land use during 1990–2010.
Fig 12.
Amount of landuse change (a) and rate of relative change of landuse (b) in different stages.
Blue and red bars respectively represent period 1990–2000 and 2000–2010. In this figure, F, G, D, P, W, B are short for forest, grass, dry farmland, paddy field, water and building lot, respectively.
Fig 13.
Land cover changes through 1990–2000 (a) and 2000–2010 (b) in study area.
F, G, D, P, W, B are short for forest, grass, dry farmland, paddy field, water and building lot, respectively. Maps in Fig 13 were generated by building land use transition matrix using ArcGIS 10.2 (the URL link is http://www.esri.com/) based on three-phase land use vector diagram which generated in Fig 11.
Table 3.
Land use transition matrix between 1990-2010(%).
Fig 14.
The time axis of influence of human activities during the research period.
Basis for division of periods are the same as Fig 10. Content in gray squares indicate the main human activities of different stages. Content in green squares show ecological phase of different stages.
Fig 15.
Double mass curve of annual precipitation and runoff during 1981–2006.
Distribution of blue triangles show cumulative values of annual precipitation and runoff. Red line indicates the regression curve line of accumulated precipitation and the accumulated runoff that establishing by linear fitting.
Fig 16.
The influence of human activities and climate change on runoff in variation section.
Yellow bars indicate the total runoff difference between the reference and variation sections. Blue and gray lines show the runoff variations caused by human activities and climate change, respectively, that were calculated according to Formulas (10) and (11). Criteria for the division of the reference section and variation section is the same as Fig 10.
Table 4.
The impact of climate change and human activities.