全 文 ：评价湖泊富营养化的一个综合模型 3
蔡庆华1 ,2 3 3 　刘建康1 　Lorenz King2
(1 中国科学院水生生物研究所 ,淡水生态与生物技术国家重点实验室 ,武汉 430072 ,中国 ;
2吉森大学地理研究所 ,吉森 ,D - 35390 ,德国)
【摘要】　湖泊富营养化的评价 ,即确定水体的状态属性 ,实际上是一个将定性问题定量化的多变量的综合
决策过程 ,因此 ,对湖泊的富营养化程度进行评价应以综合评价为主. 在综述国内外若干综合评价方法的
基础上 ,指出营养状态指数 ( TSI)法应可作为湖泊富营养化评价的主要方法 ,因其可对湖泊的营养状态进
行连续的数值化分级 ,从而为富营养化机理的定量研究提供坚实基础. 采用层次分析 (AHP)法确定综合评
价指标中的权重分配 ,构建一综合评价模型 : TS I = W (Chla) ×TS I (Chla) + W ( Sd) ×TS I ( Sd) + W ( TP)
×TS I ( TP) 或 TS IM = W (Chla) ×TS IM (Chla) + W ( Sd) ×TS IM ( Sd) + W ( TP) ×TS IM ( TP) . 此外 ,文中
简要讨论了综合评价与其他统计方法如聚类分析的关系.
关键词 　富营养化 　营养状态指数 　综合评价 　层次分析
文章编号 　1001 - 9332 (2002) 12 - 1674 - 05 　中图分类号 　X826 　文献标识码 　A
A comprehensive model for assessing lake eutrophication. Qinghua CAI1 ,2 ,Jiankang L IU1 and Lorenz King2
(1 S tate Key L aboratory of Freshw ater Ecology and Biotechnology , Institute of Hydrobiology , Chinese Academy
of Sciences , W uhan 430072 , China ;2 Institute of Geography , J ustus2L iebig2U niversity Giessen , D235390
Giessen , Germ any) . 2Chin. J . A ppl . Ecol . ,2002 ,13 (12) :1674～1678.
The evaluation of eutrophication or trophic state of a lake is in fact a multivariate comprehensive decision2making
process quantifying the qualitative problem. Therefore ,we should use a comprehensive method to assess lake eu2
trophication. On the basis of summarizing some comprehensive methods for assessing lake eutrophication reported
in China and abroad ,it is pointed out that the trophic state index ( TS I) should be a major method for evaluating
lake eutrophication , since it could provide a continuous numerical class of lake trophic state and a rigorous foun2
dation of quantitative studies of eutrophication mechanism. Using analytic hierarchy process (AHP) to determine
the weight attributions in the selected comprehensive indices , the authors ,constructed a comprehensive assess2
ment model as : TS I = W (Chla) ×TS I ( Chla) + W ( Sd) ×TS I ( Sd) + W ( TP) ×TS I ( TP) or TS IM = W
(Chla) ×TS IM (Chla) + W (Sd) ×TS IM ( Sd) + W ( TP) ×TS IM ( TP) where W ( X) were the weights for the
above three parameters with value in percentage as 54. 0 ,29. 7 and 16. 3 ,respectively. Additionally ,the relations
between comprehensive evaluation and other statistical methods such as cluster analysis were briefly discussed.
Key words 　Eutrophication , Trophic state index , Comprehensive evaluation , Analytic hierarchy process.3 The present study was supported financially by the Frontier Science
Project Programme of the Institute of Hydrobiology ,the Chinese Acade2
my of Sciences (No. 220208) ,and the National Natural Sciences Foun2
dation of China (No. 30070153 ,39670150) .3 3 Corresponding author.
Received 3 March 2002 ,Accepted 24 J une 2002.
1 　INTROD UCTION
After several decades of research describing biotic
population dynamics , relationship between popula2
tions in a community , and interaction between biotic
factors and physicochemical environmental factors ,
etc. , f reshwater ecology , especially lake ecology , has
gotten into a quite well development theoretically and
practically. From the end of World War II , man has
faced with several big problems such as overexploita2
tion of natural resources , rapid increase of human
population , and deterioration of global environment
including eutrophication of waters[15～19 ,21 ,23 ] .
Hence , the main context of f reshwater ecology re2
search is to find out the optimum water productivity ,
and to control and recover eutrophical or degenerate
waters[2 ,10 ,11 ] . These two parts are in close relation2 ship . In fact , the former is to plan the future of wa2ters and the latter is to make a summary of watersbased on the past development . The latter is often abasis of the former.Eutrophication can be understood as a phe2nomenon of the enrichment of nutrients in a waterbody due to an increase in nutrient loading. The mostimportant nutrients that cause eutrophication arephosphates , nit rates , and ammonia[4 ,8 ,13～16 ] . Themost prominent features of eutrophic waters are highcontent of nutrients and the abundance of planktonic
应 用 生 态 学 报 　2002 年 12 月 　第 13 卷 　第 12 期 　　　　　　　　　　　　　　　　　　　　　　　　　　　　　
CHIN ESE JOURNAL OF APPL IED ECOLO GY ,Dec. 2002 ,13 (12)∶1674～1678
or attached algae. There is still some discussion as to
what exactly distinguishes each trophic state. There
are also some arguments about the definition of eu2
t rophication[22 ] . OECD[18 ] defined eutrophication as
“the nutrient enrichment of waters which results in
the stimulation of an array of symptomatic changes ,
among which increased production of algae and
macrophytes , deterioration of water quality and other
symptomatic changes , are found to be undesirable as
they interfere with water uses”. This definition is ac2
cepted widely.
Many researchers have classified the trophic state
by descriptive category such as oligotrophic ,
mesotrophic , eutrophic and so on , based on the driv2
ing force or the response parameters[24 ] . All t rophic
classifications are based upon the division of the
trophic continuum , which is usually defined into a se2
ries of classes termed trophic states. Traditional sys2
tems divide the continuum into three classes : olig2
otrophic , mesotrophic , and eutrophic[9 ] . In order to
lucubrate the process of lake eutrophication and the
relationship among factors affecting lake eutrophica2
tion (such as human population growth , development
of industry and agriculture in lake catchment , etc. ) ,
it is necessary to classify lake eutrophication as con2
tinuous numerical form , so that we can undertake
further qualitative analysis[6 ] .
Some authors suggest that numerical standards
should be based on important lake parameters. One
frequently used index is the Carlson index , which is
based on chlorophyll a , Secchi depth , and total phos2
phorus[9 ] . The evaluation methods of lake eutrophica2
tion could be concluded as character method , parame2
ter method , biotic indices method , phosphorus bud2
get model method , t rophic state index method and
mathematical analysis method , etc. [5 ] . Among the
above six major methods , the trophic state index pro2
posed by Carlson[9 ] probably is the most suitable and
most acceptable method for evaluating lake eutrophi2
cation , since it provides a continuous numerical class
of lake trophic state and a rigorous foundation for
quantitative studies of eutrophication mechanism.
When we assess the eutrophic extent of certain
waters , we often encounter such a kind of problem :
selecting different parameters would lead to different
results. The reason is that evaluation of the eutrophi2
cation in (or t rophic state of) a lake is in fact a multi2
variate comprehensive decision2making process quan2
tifying the qualitative problem[8 ] . Therefore , we
should use the comprehensive method to evaluate lake
eutrophication. And the purpose of the evaluation is
to manage lake water quality , to communicate to the
public and to compare water quality in different
lakes[24 ] . Besides the establishment of a rational eval2
uation index system for the trophic state of a lake , a
key problem in this process is to determine the weight
att ributions of the selected indices.
2 　DETERMINING WEIGHT ATTRIBUTIONS IN
A COMPREHENSIVE EVAL UATION SYSTEM
　 　 The analytic hierarchy process ( AHP ) proposed by
Saaty[20 ] has been a frequently used mathematical method in
systems analysis. When we wish to compare the influences of n
elements y = { y1 , y2 ,···, yn } against object z and find their
weights , let aij as the number indicating the strength of yi
while compared with yj . The matrix of these numbers aij could
be denoted as A , or
A = ( aij) n ×n
Observably , the matrix A is a positive reciprocal one and
satisfies with :
aij > 0 , aji = 1/ aij ( i , j = 1 ,2 , ⋯,n)
If the judgment is perfect in all comparisons , then
aik ×akj = aij
for all i , j , k and Saaty[20 ] called the matrix A consistent .
If λ1 ,λ2 ,. . . ,λn (eigenvalues of A ) are the numbers satis2
fying with the equation A x =λx ,and if aii = 1 for all i , then
∑λ= n
Therefore ,in the consistent case , all eigenvalues are zero , ex2
cept one , which is n . This implies that n is the largest eigen2
value of A . If one changes the entries aij of a positive reciprocal
matrix A by small amounts , then the eigenvalues change by
small amounts.
Combining these results we found that if the diagonal of a
matrix A consisted of ones ( aii = 1) , and if A was consistent ,
then small variations of the aij kept the largest eigenvalue ,
λmax , closed to n , and the remaining eigenvalues closed to zero.
Therefore , the problem is that if A is the matrix of pairwise
comparison values , in order to find the priority vector , we must
find the vector w which satisfies with :
A w =λmax w
Since it is desirable to have a normalized solution , we altered w
slightly by settingα= ∑w i and replacing w by (1/α) w . This
ensures uniqueness , and also that ∑w i = 1.
Observing that since small changes in aij imply a small
change inλmax , the deviation of the latter from n is a measure
of consistency. It enables us to evaluate the closeness of our de2
rived scale from an underlying ratio scale which we wish to esti2
mate. Thus , we took
CI = (λmax - n) / ( n - 1)
the consistency index ( CI ) , as our indicator of“closeness to
consistency”. In general , if this number is less than 0. 1 , we
may be satisfied with our judgments[20 ] .
We shall call the consistency index of a randomly generated
reciprocal matrix from the scale 1 to 9 ,with reciprocals forced ,
the random index ( RI) . Saaty gave the order of the matrix and
the average RI . One would expect the RI to increase as the or2
576112 期 　　　　　　　　　　　　　蔡庆华等 :评价湖泊富营养化的一个综合模型 　　　　　　　　
der of the matrix increases. The ration of CI to the average RI
for the same order matrix is called the consistency ratio ( CR) .
CR = CI/ RI
A consistency ratio of 0. 10 or less is considered accept2
able[20 ] .
3 　INDEX AND ITS CRITERIA OF EVAL UATION
　　Carlson[9 ] proposed a numerical t rophic state in2
dex ( TS I) that incorporated most lakes in a scale of 0
to 100. The index number was calculated from Secchi
disk transparency ( Sd , m) , surface algal chlorophyll
a (Chla , mg·m - 3) , and total phosphorus ( TP , mg·
m
- 3) in the lake. The computational forms of the e2
quations are as follows :
TS I (Sd) = 10 ×(6 - ln (Sd) / ln2)
TS I (Chla) = 10 ×(6 - (2. 04 - 0. 68 ×(ln (Chla) ) ) / ln2)
TS I ( TP) = 10 ×(6 - ln (48/ TP) / ln2)
He also listed the completed scale and its associated
parameters in his paper [9 ] .
After having examined the possibility of the ap2
plication of TS I to Japanese lakes and the relation2
ships between this index and other parameters related
to the trophic status of lakes , Aizaki et al . [1 ] modi2
fied the Carson’s TS I as follows :
TSIM (Sd) = 10 ×(2. 46 + (3. 69 - 1. 53 ×ln(Sd) ) / ln2. 5)
TSIM (Chla) = 10 ×(2. 46 + ln(Chla) / ln2. 5)
TSIM ( TP) = 10 ×(2. 46 + (6. 71 + 1. 15 ×ln( TP) ) / ln2. 5)
The foundation of Carlson’s TS I is to assume that
the suspension in the lake consists completely of phyto2
plankton , and water transparency is affected mainly by
abundance of phytoplankton. The TS I is , therefore , cal2
culated on the base of transparency. This method ignores
the effect of other factors (including watercolor ,dissolved
matters and other suspension ,etc. ) affecting water trans2
parency[12 ]. The Aizaki’s calculation based on algal
chlorophyll a , however ,could offset the disadvantage of
Carlson’s index.
Additionally , Aizaki et al . [1 ] calculated linear
relationships between chlorophyll a and total nit rogen
( TN , mg·L - 1 ) , chemical oxygen demand ( COD ,
mg·L - 1 ) , and seston dry weight ( SS , mg·L - 1 ) ,
etc. , and produced the relevant formulation of modi2
fied TS IM . They also rebuilt the trophic state index
and its associated parameters[1 ] .
TSIM ( TN) = 10 ×(2. 46 + (3. 93 + 1. 35 ×ln( TN) ) / ln2. 5)
TSIM (COD) = 10 ×(2. 46 + (1. 50 + 1. 36 ×ln (COD) ) /
ln2. 5)
TSIM (SS) = 10 ×(2. 46 + (1. 12 + 1. 04 ×ln(SS) ) / ln2. 5)
In China , Li Zuoyong et al . [14 ] collected some
published data relating to the classification of eu2
t rophication of lakes and reservoirs in China , and
yielded the modified formula ( TS IC) for the assess2
ment of Chinese lakes and reservoirs as follows :
TS IC (Chla) = 10 ×(2. 46 + 1. 09 ×ln (Chla) )
TS IC (Sd) = 10 ×(5. 52 - 1. 94 ×ln (Sd) )
TS IC ( TP) = 10 ×(9. 40 + 1. 62 ×ln ( TP) )
TS IC ( TN) = 10 ×(5. 24 + 1. 86 ×ln ( TN) )
TS IC (COD) = 10 ×(0. 62 + 2. 56 ×ln (COD) )
TS IC (BOD) = 10 ×(2. 39 + 2. 25 ×ln (BOD) )
There are , however , some disadvantages. They
have chosen Erhai Lake in Dali , Yunnan Province ,
and Liuhuahu Lake in Guangzhou , Guangdong
Province , as typical lakes representing an oligotrophic
lake and an eutrophic one , respectively. And they
modified the formula of ( TS IC) directly f rom TS I
and TS IM instead of ascertaining from the relevant
correlation. The calculated ( TS IC) of some parame2
ters would sometimes exceed 100[5 ] .
4 　WEIGHT ATTRIBUTION
In Carlson’s TS I , the relative importance of
three factors could be considered as Chla > Sd > TP.
According to the relative importance of the trophic
state , Cai[5 ] calculated the weight assignment to the
three factors by using Saaty’s analytic hierarchy pro2
cess (AHP) . The judgment matrix (for the scoring
and meaning see Saaty[20 ]) was then presented as :
1 2 3
1/ 2 1 2
1/ 3 1/ 2 1
The weight vector is calculated as w = (0. 540 ,0. 297 ,
0. 163) , in which ,
λmax = 3. 009 , CI = 0. 005 , RI = 0. 580 , CR = 0. 006
The sum of the three weighted TS I values is called a
comprehensive TS I ( C TS I) . Cai [5 ] also classified the
trophic state of inland waters based on the C TS I ,
where , C TS I < 37 oligotrophic ; C TS I = 38 - 53
mesotrophic ; TS I > 54 eutrophic.
Saaty (1980) proposed several reasons for setting
an upper limit on the scale.
1) The qualitative distinctions are meaningful in
practice and have an element of precision when the
items being compared are of the same order of magni2
tude with regard to the property used to make the
comparison.
2) We note that our ability to make qualitative dis2
tinctions is well represented by five attributes : equal ,
weak , strong , very strong , and absolute. We can make
compromises between adjacent attributes when greater
precision is needed. The totality requires nine values and
they may well be consecutive —the resulting scale would
then be validated in practice.
6761 应 　用 　生 　态 　学 　报 　　　　　　　　　　　　　　　　　　13 卷
3) By way of reinforcing 2) , a practical method
often used to evaluate items is the classification of
stimuli into a trichotomy of regions : rejection , indif2
ference and acceptance. For finer classification , each
of these is subdivided into a trichotomy of low , medi2
um , and high2in all indicating nine shades of mean2
ingful distinctions. One of Saaty’s colleagues has in2
dicated that marketing studies conduced by his other
colleagues show that one does not need more than
about 7 scale points to distinguish between stimuli.
Thus we need not go above 9.
4) The psychological limit of 7 ±2 items in a si2
multaneous comparison suggests that if we take 7 + 2
items satisfying the description under 1) , and if they
are all slightly different f rom each other , we would
need 9 points to distinguish these differences.
Generally , we can assume the relative impor2
tance of algal chlorophyll a ( Chla) comparing with
water t ransparency (measured as Secchi disk , Sd) is
a , then , the relative importance of Sd comparing
with chla should be 1/ a . Similarly , if relative impor2
tance of Chla comparing with total phosphorus ( TP)
is b , then that of TP vs chla will be 1/ b. In the
same way , when Sd comparing with TP is c , that of
TP against Sd is , therefore , 1/ c. The judgment ma2
t rix would then be presented as :
　　
1 a b
1/ a 1 c
1/ b 1/ c 1
　　We should also suppose that a , b , and c do not
exceed 3 , since the choosy parame ( Chla , Sd , and
TP) for evaluation eutrophication are the representa2
tion of biological , physical and chemical character ,
respectively , and we could not say that one of them is
more important than others. We can , therefore , cal2
culate the weight att ribution in different sets of rela2
tive importance of Chla , Sd , and TP ( Table 1) .
　　In case of Aizaki’s modified TS I , the relative
importance among the six parameters will be :
　　Chla > Sd > TP > TN > COD > SS
The relevant judgment matrix is :
　　
1 2 3 7 8 9
1/ 2 1 2 3 5 6
1/ 3 1/ 2 1 2 3 5
1/ 7 1/ 3 1/ 2 1 2 3
1/ 8 1/ 5 1/ 3 1/ 2 1 2
1/ 9 1/ 6 1/ 5 1/ 3 1/ 2 1
The weight vector could be calculated as w = (0. 440 ,
0. 242 ,0. 149 ,0. 083 ,0. 052 ,0. 034) ,in which ,
　　λmax = 6. 083 , CI = 0. 017 , RI = 1. 240 , CR = 0. 013
Table 1 Weight attribution calculated under the different consideration
of relative importance among chlorophyll a , Secchi disk , and total
phosphorus in waters for assessing eutrophication
Parameters
a b c
Weights
W1 W2 W3 λM RI CI CR TS I TS IM
1 1 1 0. 333 0. 333 0. 333 3 0. 58 0 0. 046 54. 67 55. 00
1 1 2 0. 327 0. 413 0. 260 3. 054 0. 58 0. 027 0. 046 54. 33 54. 85
1 1 3 0. 319 0. 460 0. 221 3. 136 0. 58 0. 068 0. 117
1 2 1 0. 413 0. 327 0. 260 3. 054 0. 58 0. 027 0. 046 54. 93 54. 93
1 2 2 0. 4 0. 4 0. 2 3 0. 58 0 0 54. 60 54. 80
1 2 3 0. 387 0. 443 0. 169 3. 018 0. 58 0. 009 0. 016 54. 33 54. 67
1 3 1 0. 460 0. 319 . 0221 3. 136 0. 58 0. 068 0. 117
1 3 2 0. 443 0. 387 0. 169 3. 018 0. 58 0. 009 0. 016 54. 73 54. 73
1 3 3 0. 429 0. 429 0. 143 3 0. 58 0 0 54. 63 54. 77
2 1 1 0. 413 0. 260 0. 327 3. 054 0. 58 0. 027 0. 046 55. 20 55. 07
2 1 2 0. 413 0. 327 0. 260 3. 217 0. 58 0. 109 0. 187
2 1 3 0. 407 0. 370 0. 224 3. 367 0. 58 0. 184 0. 317
2 2 1 0. 5 0. 25 0. 25 3 0. 58 0 0 55. 50 55. 00
2 2 2 0. 493 0. 311 0. 196 3. 054 0. 58 0. 027 0. 046 55. 24 54. 89
2 2 3 0. 484 0. 349 0. 168 3. 136 0. 58 0. 068 0. 117
2 3 1 0. 550 0. 240 0. 210 3. 018 0. 58 0. 009 0. 016 55. 69 54. 97
2 3 2 0. 540 0. 297 0. 163 3. 009 0. 58 0. 005 0. 008 55. 43 54. 87
2 3 3 0. 528 0. 333 0. 140 3. 054 0. 58 0. 027 0. 046 55. 31 54. 86
3 1 1 0. 460 0. 221 0. 319 3. 136 0. 58 0. 068 0. 117
3 1 2 0. 464 0. 281 0. 255 3. 367 0. 58 0. 184 0. 317
3 1 3 0. 460 0. 319 0. 221 3. 561 0. 58 0. 280 0. 483
3 2 1 0. 55 0. 21 0. 24 3. 018 0. 58 0. 009 0. 016 55. 81 55. 03
3 2 2 0. 547 0. 263 0. 190 3. 136 0. 58 0. 068 0. 117
3 2 3 0. 54 0. 297 0. 163 3. 257 0. 58 0. 128 0. 221
3 3 1 0. 6 0. 2 0. 2 3 0. 58 0 0 56. 00 55. 00
3 3 2 0. 594 0. 249 0. 157 3. 054 0. 58 0. 027 0. 046 55. 79 54. 91
3 3 3 0. 584 0. 281 0. 135 3. 136 0. 58 0. 068 0. 117
Indicate a satisfactory consistency. The results
showed that the above weight att ribution could repre2
sent well the relative importance of several parame2
ters ,and could be used for assessing lake eutrophica2
tion.
5 　RELATIONS BETWEEN COMPREHENSIVE E2
VAL UATION AND OTHER STATISTICAL METH2
ODS SUCHAS CL USTER ANALYSIS
　　The authors , therefore , built a comprehensive
assessment model after calculating the weight att ribu2
tion of the three TS I (and TS IM ) indices by using
the method of analytic hierarchy process. The model
was listed as :
　　TS I = W (Chla) ×TS I (Chla) + W (Sd) ×TS I (Sd) + W
　　　( TP) ×TS I ( TP)
or
TS IM = W (Chla) ×TS IM (Chla) + W (Sd) ×TS IM (Sd)
　　　+ W ( TP) ×TS IM ( TP)
where W ( X ) were the weights for the above three
parameters with value in percentage as 54. 0 , 29. 7
and 16. 3 , respectively. The right two columns of
Table 1 are the corresponding evaluation of Lake
Donghu in Wuhan , China , by Carlson’s TS I and
Aizaki’s TS IM ( shown only the case of significance ,
CR < 0. 10) [4 ] .
　　Comprehensive evaluation is , in fact , a mapping
776112 期 　　　　　　　　　　　　　蔡庆华等 :评价湖泊富营养化的一个综合模型 　　　　　　　　
process through which the points of n2dimensional
space consisted of n indices are projected into 12di2
mensional space , to get an easy comparable numer2
al [8 ] . Therefore , there will be more or less loss of in2
formation during projection. The cluster analysis ,
however , deals directly with the correlation among
points in the n2dimensional space , and classifies them
into groups according their correlation[3 ] . It , then ,
avoids loss of information during projection , but could
not compare qualitatively the magnitude of groups. It
is very necessary combining the two methods to get
more objective and accurate result when assessing
trophic states of waters.
　　Cai et al . [7 ] evaluated the environmental quality
of some lakes near the Chinese Zhongshan Station and
Great Wall Station , Antarctica , based on the trophic
state index ( Fig. 1 and Table 2) . The result indicat2
ed that these lakes were oligotrophic or mesotrophic
waters , especially Mochou Lake at Zhongshan Station
and Xihu Lake at Great Wall Station , where the lake
water was of good quality and suitable for drinking.
Comparing the results of Xihu Lake between 1995
and 1993 , however , it indicated that the water quali2
ty of the lake declined.
Fig. 1 Cluster dendrogram of trophic state for some lakes near Chinese
Zhongshan Station and Great Wall Station , Antarctica.
1) Gaoshan L . ;2) Xihu L . (’95) ;3) Tern L . ;4) Xihu L . (’93) ;5) Mo2
chou L .
Table 2 Comprehensive evaluation for trophic state of some lakes near
the Chinese Zhongshan and Great Wall Stations , Antarctica
Name
of
lakes
Samp2
ling
year
Chla
μg·
L - 1
TSI
　
TS IM
　
Sd
m
　
TS I
　
TS IM
　
TP
mg·
L - 1
TS I
　
TS IM
　
Weighted mean
TS I
　
TS IM
　
Mochou 1992 0. 36 21 13 3. 3 43 45 0. 026 51 52 32 29
Gaoshan 1993 1. 39 34 28 2. 6 46 49 0. 035 55 56 41 39
Tern 1993 0. 91 30 24 2. 5 47 50 0. 040 57 57 39 37
Xihu 1993 0. 50 24 17 2. 7 46 48 0. 024 50 51 35 32
Xihu 1995 1. 29 33 27 2. 7 46 48 0. 030 53 54 40 38
Acknowledgements
　　The Max Planck Society (MPG) provided the opportunity
to exchange scientists between the Institute of Hydrobiology
( IHB) , Chinese Academy of Sciences (CAS) , and the Institute
of Geography of J ustus2Liebig2University Giessen , Germany.
References
1 　Aizaki M , et al . 1981. Application of modified Carlson’s trophic
state index to Japanese lakes and its relationships to other parame2
ters related to trophic state. Res Rep Nat Inst Envi ron S t udy ,23 :
13～31 (in Japanese ,with English abstract) .
2 　Allen JD ,Johnson LB. 1997. Catchment2scale analysis of aquatic e2
cosystem. Freshw ater Biol ,37 :107～111
3 　Cai Q. 1988. Fuzzy cluster analysis of pollution in Lake Donghu e2
cosystem. Acta Hydrobiol S in ,12 (3) :193～198 (in Chinese ,with
English abstract) .
4 　Cai Q. 1993. Comprehensive evaluation of eutrophication in Donghu
Lake ,Wuhan. Oceanol L i m nol S in ,24 (4) :335～339 (in Chinese ,
with English abstract) .
5 　Cai Q. 1997. On the comprehensive evaluation methods for lake eu2
trophication. J L ake Sci ,9 (1) : 89～94 (in Chinese , with English
abstract) .
6 　Cai Q ,Liu J . 1994. Effect of human population growth and fishery
development on water quality of Lake Donghu , Wuhan. Acta Hy2
drobiol S in ,18 (1) :87～89 (in Chinese ,with English abstract) .
7 　Cai Q ,Wang J ,Li Z , et al . 1997. Evaluation of environmental quali2
ty of some lakes around Chinese Zhongshan Station and Great Wall
Station ,Antarctica , based on trophic state index. Oceanol L i m nol
S in ,28 (supp . ) :134～138 (in Chinese ,with English abstract) .
8 　Cai Q ,Liu R ,Li D , et al . 1992. Comprehensive evaluation of troph2
ic states of water systems in Fengqiu Experimental Area of Huang2
Huai2Hai Plain. J L ake Sci ,4 (2) :46～51 (in Chinese ,with Eng2
lish abstract) .
9 　Carlson RE. 1977. A trophic state index for lakes. L i m nology and
Oceanography ,22 (2) :361～369
10 　Chen J , Gao X , He D , et al . 2000. Nitrogen contamination in the
Yangtze River system , China. J Hazardous M aterials A , 73 ( 1) :
107～113
11 　Engel DW , Thayer GW , Evans DW. 1999. Linkages between fish2
ery habitat quality , stressors , and fishery population. Envi ron Sci
Policy ,2 :465～475.
12 Goda T. 1981. Comprehensive studies on the eutrophication of
fresh2water areas Ⅺ. Summary of researches. Nat Inst Envi ron
S t udies ,27 :59～71 (in Japanese ,with English abstract) .
13 　Horne AJ and Goldman CR. 1994. Limnology. 2nd ed. New York :
Mc Graw2Hill Inc.
14 　Li Z and Zhang H. 1993. Trophic state index and its correlation
with lake parameters. Acta Sci Ci rc ,13 (4) : 391～397 (in Chi2
nese ,with English abstract)
15 Lijklema L . 1995. Development and eutrophication : experiences
and perspectives. W ater Sci Technol ,31 (9) :11～15
16 　Liu J . 1984. Lakes of the middle and lower basins of Changjiang
(China) . In : Taub FB ed. Lake and Reservoirs. Amsterdam : Elsevi2
er. pp331～355
17 　Jin X ,Liu H , Tu Q , et al . 1990. Eutrophication of lakes in China.
Beijing : China Environmental Sciences Press. (in Chinese)
18 　OECD. 1982. Eutrophication of waters : monitoring ,assessment and
control. Paris :OECD.
19 　Ryding SO , Rast W. 1989. The control of eutrophication of lakes
and reservoirs. Paris : The Parthenon Publishing Group .
20 　Saaty TL . 1980. The analytic hierarchy process. New York : Mc2
Graw2Hill Inc.
21 　Smith VH , Tilman GD and Nekola J C. 1999. Eutrophication : im2
pacts of excess nutrient inputs on freshwater ,marine ,and terrestrial
ecosystems. Envi ron Poll ,100 :179～196
22 　Wetzel RG. 1983. Limnology. 2nd ed. Philadelphia : Saunders Col2
lege Publ. Co.
23 　Yang MD ,Sykes RM and Merry CJ . 2000. Estimation of algal bio2
logical parameters using water quality modeling and SPOT satellite
data. Ecol Mod ,125 (1) :1～13
24 Yoshimi H. 1987. Simultaneous construction of single2parameter
and multiparameter trophic state indices. W ater Res ,21 (12) :1505
～1511
作者简介 　蔡庆华 ,男 ,1964 年生 ,博士 ,研究员 ,主要从事
淡水生态学、流域生态学和系统生态学研究 ,发表学术论文
80 余篇. E2mail : qhcai @ihb. ac. cn.
8761 应 　用 　生 　态 　学 　报 　　　　　　　　　　　　　　　　　　13 卷