Agriculture Reference
In-Depth Information
900 ha, which is far smaller than the average area of the Beijing mountain
towns (9,285 ha), was determined as the minimum number of cells for a
stream network to delineate basic watershed units (Figure 3).
1.2.3.2 DELINEATING MULTI-SCALE WATERSHEDS AS
WATERSHED ANALYSIS UNITS
Multi-scale watersheds corresponding to the 27 monitoring sites were
identified as watershed analysis units for further analysis according to the
flow direction and rivers, respectively. Since some monitoring sites are
located on the same streams, such as Sites 1-3 (Bai River), Sites 18-20
(Qingshui River) and Sites 21-25 (Yongding River), perhaps they are sta-
tistically highly correlated and, to some extent, all the upstream points
contribute to the measurements of any monitoring point. This is statis-
tically undesirable and would produce strongly biased results. To solve
this problem, the independence of these sites' data should be tested. A
serial autocorrelation test was adopted to analyze the possible correlation
between neighboring sites. The test results showed that these upstream
monitoring sites had little contribution to their nearest downstream sites
for NO
3
-N and COD
Mn
water quality data in 2000. Therefore, the monitor-
ing samples can be considered independent for further statistical analysis,
and the watersheds contributing flow to these sites cannot be included in
the their nearest downstream sites' watersheds when delineating the wa-
tershed analysis units.
Firstly, the basic watershed units contributing fl ow to every monitoring
site were delineated as the whole watershed of every monitoring site, which
did not have a nest relation according to the result of serial autocorrela-
tion test mentioned above (Figure 4). Table 3 shows the number of towns
covered by the whole watershed for every monitoring site. Subsequently,
multi-scale watersheds were determined in the whole watershed area of
every monitoring site. The basic watershed units directly contributing fl ow
to every monitoring site were considered as the fi rst order watershed (Zone
1). The Zone 1 and the basic watershed units directly contributing fl ow to
the Zone 1 were together defi ned as second order watershed (Zone 2).
