Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, Vol. 4, Iss. 3
Transcript of Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, Vol. 4, Iss. 3
Proceedings of the International Academy of
Ecology and Environmental Sciences
Vol. 4, No. 3, 1 September 2014
International Academy of Ecology and Environmental Sciences
Proceedings of the International Academy of Ecology and Environmental Sciences ISSN 2220-8860 Volume 4, Number 3, 1 September2014
Editor-in-Chief WenJun Zhang Sun Yat-sen University, China International Academy of Ecology and Environmental Sciences, Hong Kong E-mail: [email protected], [email protected] Editorial Board Taicheng An (Guangzhou Institute of Geochemistry, Chinese Academy of Sciences, China) Jayanath Ananda (La Trobe University, Australia) Ronaldo Angelini (The Federal University of Rio Grande do Norte, Brazil) Nabin Baral (Virginia Polytechnic Institute and State University, USA) Andre Bianconi (Sao Paulo State University (Unesp), Brazil) Iris Bohnet (CSIRO, James Cook University, Australia) Goutam Chandra (Burdwan University, India) Daniela Cianelli (University of Naples Parthenope, Italy) Alessandro Ferrarini (University of Parma, Italy) Marcello Iriti (Milan State University, Italy) Vladimir Krivtsov (Heriot-Watt University, UK) Suyash Kumar (Govt. PG Science College, India) Frank Lemckert (Industry and Investment NSW, Australia) Bryan F. J. Manly (Western EcoSystems Technology Inc. and University of Wyoming, USA) T.N. Manohara (Rain Forest Research Institute, India) Ioannis M. Meliadis (Forest Research Institute, Greece) Lev V. Nedorezov (University of Nova Gorica, Slovenia) George P. Petropoulos (Institute of Applied and Computational Mathematics, Greece) Edoardo Puglisi (Università Cattolica del Sacro Cuore, Italy) Zeyuan Qiu (New Jersey Institute of Technology, USA) Mohammad Hossein Sayadi Anari (University of Birjand, Iran) Mohammed Rafi G. Sayyed (Poona College, India) R.N. Tiwari (Govt. P.G.Science College, India) Editorial Office: [email protected] Publisher: International Academy of Ecology and Environmental Sciences Address: Flat C, 23/F, Lucky Plaza, 315-321 Lockhart Road, Wanchai, Hong Kong Tel: 00852-6555 7188; Fax: 00852-3177 9906 Website: http://www.iaees.org/ E-mail: [email protected]
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
Article
Morphometry and meristic counts of Bombay duck, Harpodon
nehereus (Hamilton, 1822) along Sunderban region of West Bengal,
India
V. Vinaya Kumar1,4, A. Devivaraprasad Reddy2, Sampurna Roy Choudhury1, C. H. Balakrishna3, Y. Satyanaryana3, T.S. Nagesh1, Sudhir Kumar Das1 1Department of Fishery Biology and Resources Management, Faculty of Fishery Sciences, WBUAFS, Kolkata – 94, West Bengal,
India 2Department of Fish Processing Technology, College of Fishery Science, SVVU, Muthukur-524 344, Andhra Pradesh, India 3Fisheries Development Officer, Joint Director Fisheries, Fishing Harbor, Vishakapatnam, Andhra Pradesh, India 4Fishery Resources Assessment Division, Central Marine Fisheries Research Institute, Cohin, Kerala, India
E-mail: [email protected]
Received 4 January 2014; Accepted 10 February 2014; Published online 1 September 2014
Abstract
Fisheries sector have been gaining importance globally due to their role in national economy, foreign exchange
earnings and employment generation besides providing nutritious food and cheap protein not only to the fisher
folk but also to the rapidly growing population. Bombay duck fishery supported by single species, Harpodon
nehereus, contributes about 4-5 % of the estimated average annual marine landings of India. With a peculiar
discontinuous distribution fishery is utmost importance in two maritime states of India i.e. Gujarat and
Maharashtra contributing 92% of the total landings and the remaining 8% landings were from West Bengal
and Orissa coasts. H. nehereus forms a commercial fishery along Hooghly estuarine systems. The present
study aims on the morphometric and meristic counts of H. nehereus. During the period of investigation, 373
fish samples with length range (145 to 302 mm) and weight range (28 to 212 gm) were examined. Highest
significant correlation (P<0.01) was observed between reference length and other morphometric parameters of
both sexes. Percentage range difference in male's morphometric characters like post orbital length (15.24) and
snout length (15.04) are environmentally controlled and others like standard length (11.09), pre-dorsal length
(12.18), height of pelvic fin (13.39) and height of pectoral fin (12.10) are intermediate controlled (genetic and
environmental factors). But in case of females, none of the characters are controlled by environmental factors
and parameters like pre-dorsal length (10.37) and post orbital length (12.37) are intermediate controlled,
remaining parameters in both sexes are genetically controlled (hereditary). Meristic counts includes dorsal fin
with 10-13 soft rays, pelvic fin with 9 soft rays, pectoral fin with 10-12 soft rays and anal fin with 13-15 soft
rays.
Keywords Bombay duck; Harpodon nehereus; morphometry; meristic counts; Sunderban region.
1 Introduction
Proceedings of the International Academy of Ecology and Environmental Sciences ISSN 22208860 URL: http://www.iaees.org/publications/journals/piaees/onlineversion.asp RSS: http://www.iaees.org/publications/journals/piaees/rss.xml Email: [email protected] EditorinChief: WenJun Zhang Publisher: International Academy of Ecology and Environmental Sciences
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
1 Introduction
As we are proceeding in this millennium, finfish and other aquatic products will be in acute short supply as
domestic and International demand for both high and low valued species increasing due to raising populations,
living standards and disposable incomes. Bombay duck, Harpodon nehereus is a key contributor in Indian
marine fish landings ranging from 4-5% commonly along North-West and North-East coast (Fig. 1). Bombay
duck production was 1, 15,296 tonnes in 2012, contributing nearly 3-4% of the total marine landings of India
(CMFRI Annual Report, 2013). Though well relished and considered a delicacy in Western India, its culinary
qualities have not been recognized in West Bengal. H. nehereus forms a lucrative fishery along Sunderban
region of North-East coast of India. Bombay duck is a very soft and highly perishable due to high moisture
content in its muscle. It is having good importance and relished by different sections of people as table fish and
also valuable as laminated or dried from (Kumar et al., 2012a).
Fig. 1 Sampling site i.e. Diamond harbor (22o 12’53.92’’ N & 88o 12’22.74’’ E)
Fig. 2 A view of Harpodon nehereus during measuring of morphometric & meristic parameters.
96
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
The fishery, biology and population characteristics of the H. nehereus from the Saurashtra coast were
extensively studies (Balli et al., 2011; Ghosh et al., 2009; Bapat, 1970; Khan, 1985, 1986a, 1986b, 1987 and
1989). Along Hooghly matlah estuarine region of North-East cast of India, food and feeding habits was studied
(Pillay, 1951, 1953; Kumar et al., 2012b)and population dynamics was also estimated (Krishnayya, 1968).
Studies on the morphometry and meristic counts are vital for the differentiation of taxonomic units. Studies on
variation in morphological characters are critical in order to elucidate patterns observed in phenotypic and
genotypic variations among coastal fish populations (Beheregaray and Levy, 2000). There were no studies
related with morphometry and meristic counts of Bombay duck, H. nehereus stocks along North-East coast of
India. The present work aims to full fill the research gap, upgrade the biological information of species and
also study the factors which influence the stock dynamics.
2 Materials and Methods
2.1 Sampling site and size
The present work aims on some aspects of morphometric and meristic characters of H. nehereus for the period
of one year (August, 2008 to July, 2009). The samples were collected from the Daimond harbour area (22o
12’53.92’’ N & 88o 12’22.74’’ E), Sagar Islands, Bokkhali, 8-Jetighat and local fish markets, which were
mainly procured from different areas of Sundarban region of South 24 – Paraganas district (Fig. 1). Samples
were captured mostly by stationary bag-net, locally called Beenjal, Behundijal which are non-selective
multispecies small meshed nets.
Current experiment, total of 373 specimens of H. nehereus was sampled for the 12 months period (August,
2008 to July, 2009). More than 30 specimens were examined in the laboratory during each month. Samples
were collected twice in a month and examined usually at fortnightly intervals. Total length and standard length
were measured in market itself by using the millimeter scale (Fig. 2). Total weight was measured with a
monopan balance for individual fish in grams.
2.2 Sampling method
For study of the morphometric and meristic characters the standard procedure (Lowe-McConnel, 1971) was
followed. All linear measurements were rounded to the nearest mm. Among different morphometric characters,
standard length, head length, pre-dorsal length, length base of dorsal fin, length base of anal fin, pectoral fin
length, dorsal height, pectoral fin height, least depth of caudal peduncle, post orbital length, snout length, eye
diameter were measured. Four meristic characters such as dorsal fin rays, pectoral fin rays, pelvic fin rays and
anal fin rays were estimated. Total length and head length were used as reference length. Total length was
measured from tip of the snout to the tip of the caudal fin. The diameter of the eye was measured in horizontal
axis. The regression of various morphometric characters on standard length was obtained by least square
method with the formula Y = a + bX, where, ‘Y’ is different morphometric measurements and ‘X’ is the
reference length; ‘a’ is the constant value; ‘b’ is the exponent.
2.3 Statistical analysis
Correlation co-efficient between variables were calculated and regression equations were found out following
standard methods. Isometric growth was tested by employing Fisher’s t test. Significant difference among
mean of different biological parameters were tested employing standard statistical tools like Student’s‘t’ test,
and ANOVA (Snedecor and Cochran, 1967), etc.
3 Results and Discussion
Studies on the morphometric and meristic characters of fishes provide substantial information with regard to
the exact nature of stocks and their geographical distributions. Morphometric differences are seen with in the
97
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
species and even within different sexes of species due to interactive genetic and environmental effects. The
knowledge of exact genetically and environmental controlling characters are essential for the identification of
species of a genus and the populations with in a species. Current study reveals that the biometry values H.
nehereus showed a proportional positive increase with total length of fish. The mean, percentage range,
percentage difference and standard error values of different morphological characters of H. nehereus are
presented in Fig. 3, 4 and Table 1, 2.
The regression coefficient ‘b’ of different variable characters (Yi) on the total length (X) was highest in
case of standard length. The ‘b’ values for standard length on total length are 0.5747 for male and 0.9067 for
females respectively. Least depth of caudal peduncle (Y) on total length (X), showing lowest ‘b’ value i.e.
0.0281 for males and 0.0389 for females. While considering the post orbital length, snout length and eye
diameter in terms of percentage of head length, post orbital length shows highest ‘b’ values, which were
0.7744 for males and 0.8015 for females.
The present study revealed the highest correlation of standard length on total length in case of both male
(r =0.965) and female (r =0.961) and also observed the highest correlation of post orbital length on head length
in case of both male (r =0.948) and female (r =0.957). The lowest value of the correlation for the male was
noticed in the case of height of pelvic fin (r =0.771) on total length and snout length (r =0.642) on percentage
of head length. In female lowest correlation value was observed in case of length base of anal fin (r =0.597) on
total length and snout length (r =0.572) on percentage of head length. Morphometric analysis of the present
study revealed that the correlation values were greater in male than female when calculating the percentage on
total length and head length. Only the post orbital length giving the more value in case of female (r =0.957)
when compare to male (r =0.948). All the regression coefficient ‘b’ values, correlation differentiation ‘R2’ and
coefficient of correlation ‘r’ values are represented in Table 3.
Nikolsky (1963) stated that the males and females often differ in the length and shapes of fins. Phenotypic
plasticity during present investigation occurred due to the environmental factors because fishes were procured
from different water bodies of Sunderban areas of Hooghly-Matlah estuarine system. Johal et al. (1994),
classified three categories of morphometric characters based on percentage range difference i.e. genetically
controlled characters (<10% range difference), intermediate (10.1% to 14-99% range difference) and
environmentally controlled characters (>15% percentage difference). In current study, in male parameters like
post orbital length (15.24%) and snout length (15.04%) are environmentally controlled and the other
parameters like standard length (10.09%), pre-dorsal length (12.18%), height of pelvic fin (13.39%) and height
of pectoral fin (12.10%) were controlled by intermediate factors, but in case of female pre-dorsal length
(10.37%) and post orbital length (12.37%)were controlled by intermediate factors. Other than these parameters,
all the remaining parameters in both male and female were controlled by genetic factors (hereditary).
Meristic characters of H. nehereus (Table 4) in the current study includes dorsal fin with 10-13 soft rays,
pelvic fin with 9 soft rays, pectoral fin with 10-12 soft rays and the anal fin with 13-15 soft rays. The
variations in the number of meristic characters have been documented by many workers (Abdurahiman et al.,
2004), who opined that the environmental factors particularly that the temperature influences meristic
characters in the process of their growth in fishes. The variations can also be exhibited by various stocks found
in different geographical areas (Sarker et al., 2004).The present study agreed with the previous work (Bapatet
al., 1970). The meristic counts in both of the sexes were found to be quite similar resembling the earlier work.
98
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
99
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
Fig. 3 Morphometric analysis of Bombayduck, Harpodon nehereus in Male.
100
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
Fig. 4 Morphometric analysis of Bombayduck, Harpodon nehereus in Female.
101
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
Table 1 Morphometric analysis of Harpodon nehereus (Male).
Table 2 Morphometric analysis of Harpodon nehereus (Female).
Parameters
Mean
SE
Percentage Range (min-max)
Percentage range difference
% on Total length
Standard length 83.685 0.146 78.07 - 89.16 11.09
Head length 17.390 0.083 20.94 - 13.6 7.34
Pre-dorsal length 37.796 0.128 30.26 - 42.44 12.18 Length base of Dorsal fin 12.766 0.057 10.96 - 16.06 5.13 Length base of Anal fin 14.250 0.065 11.83 - 17.25 5.42
Height of Dorsal fin 16.474 0.074 13.22 - 19.17 5.95 Height of Pelvic fin 25.510 0.125 18.61 – 32.00 13.39 Height of Pectoral fin 24.266 0.158 18.18 - 30.28 12.1
Least depth of Caudal fin 4.693 0.021 4.00 - 5.92 1.92
% on Head length
Post-orbital length 79.575 0.210 73.33 - 88.57 15.24 Snout length 20.693 0.163 13.08 - 28.12 15.04
Eye diameter 12.487 0.090 10.00 - 18.52 8.52
Parameters
Mean
SE
Percentage Range (min-max)
Percentage range difference
% on Total length
Standard length 84.304 0.191 80.00 - 89.41 9.41 Head length 17.751 0.129 14.28 - 20.82 6.54 Pre-dorsal length 37.986 0.176 31.09 - 41.46 10.37 Length base of Dorsal fin 12.436 0.079 10.60 - 14.53 3.93 Length base of Anal fin 13.634 0.105 11.06 - 17.09 6.03 Height of Dorsal fin 15.805 0.103 12.36 - 17.95 5.59 Height of Pelvic fin 24.447 0.153 20.00 - 28.51 8.51 Height of Pectoral fin 23.347 0.147 19.20 - 27.87 8.67 Least depth of Caudal fin 4.679 0.030 4.02 - 5.53 1.51 % on Head length
Post-orbital length 79.453 0.229 75.51 - 87.88 12.37 Snout length 19.273 0.197 14.63 - 24.24 9.61 Eye diameter 12.227 0.112 9.75 - 15.15 5.4
102
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
Table 3 Regression equation of morphometric parameters of Harpodon nehereus.
Table 4 Meristic characters of Harpodon nehereus.
Parameters
Male Female
Regression equation
R2 r Regression equation
R2 r
Standard length (Y) on Total length (X)
Y=5.6675+0.5747x
0.932
0.965
Y= -1.5976+0.9067x
0.922
0.961
Head length (Y) on Total length (X)
Y=0.7397+0.1406x
0.746
0.864
Y= -1.5685+0.2398x
0.657
0.811
Pre-dorsal length (Y) on Total length (X)
Y=2.2972+0.2728x
0.831
0.912
Y= -0.5698+0.4025x
0.711
0.844
Length base of Dorsal fin (Y) on Total length (X)
Y=1.0956+0.0765x
0.661
0.813
Y=0.3788+0.1093x
0.530
0.729
Length base of Anal fin (Y) on Total length (X)
Y=1.5891+0.068x
0.589
0.773
Y=0.7489+0.1066x
0.356
0.597
Height of Dorsal fin (Y) on Total length (X)
Y=1.5413+0.0927x
0.649
0.806
Y=1.2706+0.1079x
0.394
0.628
Height of Pelvic fin (Y) on Total length (X)
Y=2.4058+0.142x
0.594
0.771
Y=1.862+0.1714x
0.440
0.663
Height of Pectoral fin (Y) on Total length (X)
Y=1.9848+0.1501x
0.607
0.779
Y=1.8812+0.1611x
0.415
0.643
Least depth of Caudal fin (Y) on Total length (X)
Y=0.4004+0.0281x
0.628
0.793
Y=0.1777+0.0389x
0.334
0.578
Post-orbital length (Y) on Head length (X)
Y=0.0602+0.7744x
0.898
0.948
Y=0.064+0.8015x
0.916
0.957
Snout length (Y) on Head length (X)
Y=0.3194+0.1204x
0.412
0.642
Y= 0.4097+0.1028x
0.327
0.572
Eye diameter (Y) on Head length (X)
Y=0.149+0.0832x
0.436
0.661
Y=0.1258+0.0917x
0.394
0.628
Parameter Meristic counts
Number of Dorsal fin rays 11 – 13 soft rays
Number of Pectoral fin rays 10 – 12 soft rays
Number of Pelvic fin rays 9 soft rays
Number of Anal fin rays 13 – 15 soft rays
103
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
4 Conclusion
Current study, shows high significant correlation (P<0.01) between reference length and other morphometric
features of the both sexes. On the percentage of range difference in case of male the morphometric characters
like Post orbital length and Snout length are environmentally controlled and the other characters like standard
length, pre-dorsal length, height of pelvic fin and height of pectoral fin are intermediate controlled. But, in case
of female none of morphometric characters are controlled by environmental factors and the parameters like
pre-dorsal length and post orbital length are intermediate controlled. However, the results clearly reveal that
the biometry values of H. nehereus showed a proportional positive increase with total length of fish.
References
Abdurahiman KP, Harishnayak T, Zacharia PU, Mohamed KS. 2004. Length-weight relationship of
commercially important marine fishes and shell fishes of the southern cost of Karnataka, India. World Fish
Centre Quarterly, 27(1): 9-10
Balli JJ, Chakraborthy SK, Jaiswar AK. 2011. Population dynamics of bombay duck Harpadontidaeneherus
(Ham. 1822) (Teleostomi/Harpadontidae) from Mumbai water, India. Indian Journal of Geo-Marine
Science, 40(1): 67-70
Bapat SV. 1970. The Bombay duck, Harpodon nehereus (Ham). Bulletin CMFRI, 21: 1-66
Beheregaray LB, Levy JA. 2000. Population genetics of the silverside Odontesthesargentinensis (Teleostei,
Atherinopsidae): evidence for speciation in an estuary of southern Brazil. Capeio, 441-447
CMFRI Annual Report, 2013. Central Marine Fisheries Research Institute, Cochin, Kerala, India
Ghosh S, Pillai NGK, Dhokia HK. 2009. Fishery and population dynamics of Harpadonnehereus (Ham.) off
the Saurashtra coast. Indian Journal of Fisheries, 56(1): 13-19
Johal MS, Tandon KK, Sandhu GS. 1994. Mahseer in lacustrine water, Gobindsagar reservoir. In: Mahseer –
the Game Fish (garhwal) (Nautiyal P, ed). 65-85, JagdambePrakashan, Dehradun, India
Khan MZ. 1987. A note on the Dol net fishery off Jaffrabad (Gujarat) with special reference to Bombay duck,
from 1979-‘80 to 1981-’82. Indian Journal of Fisheries, 34(2): 188-192
Khan MZ. 1986a. Dol net fishery off Nawabunder (Gujarat). Fishery Technology, 23: 92-99
Khan MZ. 1986b. Mortality and stock size estimates of the Bombay duck, Harpodon nehereus (Ham.) off
Nawabunder (Gujarat). Indian Journal of Fisheries, 33(3): 354-358
Khan MZ. 1985. Observations on the fishery of Bombay duck,Harpodon nehereus (Ham.) along the
Saurashstra coast. Indian Journal of Fisheries, 32 (4): 431-438
Khan MZ. 1989. Population dynamics of the Bombay duck, Harpodon nehereus (Ham.) off Saurashstra coast.
Indian Journal of Fisheries, 36 (2): 93-101
Krishnayya CH. 1968. Age and growth of Harpodon nehereus (Ham) and its fishery in the Hooghly estuarine
system. Journal of Zoological Society of India, 20(1&2): 129-147
Kumar VV, Reddy AD, Balakrishna C, et al. 2012a. Ecological parameters suitable for Bombayduck fishery,
Harpodon nehereus along Sunderban area of West Bengal, India. Environment and Ecology, 30(3B): 860-
864
Kumar VV, Reddy AD, Balakrishna C, et al. 2012b. Analysis of Diet composition, feeding dynamics and
proximate composition of Bombayduck, Harpodon nehereus along Sunderban area of West Bengal, India.
Achieves in Applied Science Research, 4(2): 1175-1182
104
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 95-105
IAEES www.iaees.org
Lowe-McConnel RH. 1971. Identification of Freshwater fishes. In: Methods for Assessment of Fish
Production in Freshwater (Ricker WE, ed). 450-489, Black well Scientific Publication, Edinburg, Oxford,
UK
Nikolsky GV. 1963. The Ecology of Fishes. Academic Press, New York, USA
Pillay TVR. 1951. A preliminary note on the food and feeding habits of the Bombay duck, Harpodon nehereus
(Ham.) in the river Matlah. Science and Culture, 17: 261-262
Pillay TVR. 1953. The food and feeding habits of the Bombay duck, Harpodon nehereus (Ham.) in the river
Matlah (Bengal). Proceedings of National Institute of Science in India, 19: 427-435
Sarker Y, Jaiswar AK, Chakraborthy SK, et al. 2004. Morphometry and length weight relationship of
Megalaspiscordyla (Linnaeus, 1758) from Mumbai coast. Indian Journal of Fisheries, 51(4): 481-486
Snedecor GW, Cochran WG. 1967. Statistical Methods. Oxford and IBH Publishing Co., Calcutta, India
105
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
Article
Classic models of population dynamics: assumptions about self-
regulative mechanisms and numbers of interactions between
individuals
L.V. Nedorezov
University of Nova Gorica, Vipavska Cesta 13, Nova Gorica SI-5000, Slovenia
E-mail: [email protected]
Received 13 May 2014; Accepted 15 June 2014; Published online 1 September 2014
Abstract
Stochastic model of migrations of individuals within the limits of finite domain on a plane is considered. It is
assumed that population size scale is homogeneous, and there doesn’t exist an interval of optimal values of
population size (Alley effect doesn’t realize for population). For every fixed value of population size number
of interactions between individuals is calculated (as average in space and time). Correspondence between
several classic models and numbers of interactions between individuals is analyzed.
Keywords stochastic models; migrations; mechanistic models; self-regulative mechanisms.
1 Introduction
1 Introduction
Verhulst model (Verhulst, 1838) is one of the basic models in ecological modeling:
2xxdt
dx . (1)
In (1) )(tx is population size (or population density) at time t ; parameter is equal to difference between
intensity of birth rate and intensity of death rate; parameter , 0 const , is coefficient of influence of
self-regulative mechanisms on population dynamics; parameter /K (when 0 , and population
doesn’t eliminate for all initial values of population size) is maximum of population size which can be
achieved asymptotically. This is standard explanation of biological sense of model (1) parameters.
In (1) it is assumed that increasing of influence of self-regulative mechanisms on population size
changing (and, respectively, increasing of death rate) is proportional to population size squared (or population
density squared). This assumption is based on physical idea about paired interactions between physical objects.
In other words, it is assumed that number of interactions between individuals during rather short time period
Proceedings of the International Academy of Ecology and Environmental Sciences ISSN 22208860 URL: http://www.iaees.org/publications/journals/piaees/onlineversion.asp RSS: http://www.iaees.org/publications/journals/piaees/rss.xml Email: [email protected] EditorinChief: WenJun Zhang Publisher: International Academy of Ecology and Environmental Sciences
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
0t is equal to tx 2 . Increasing of number of interactions leads, for example, to increase of intra-
population competition for food and space, to increase of speed of a spread of diseases in population and so on.
Thus, it leads to increase of influence of self-regulative mechanisms on population dynamics. But we have
some differences between physical and biological objects… Moreover, it isn’t obvious how we have to
determine “number of interactions” for individuals even in most primitive cases (when space is homogenous,
Alley effect doesn’t realize for considering population etc.; Allee, 1931; Odum, 1983).
Comparison of theoretical results obtained with model (1) with empirical and experimental time series
showed that in various cases this model doesn’t allow obtaining good fitting for existing datasets (see, for
example, Gause, 1934; Maynard, 1968, 1974; Pielou, 1977; Isaev et al., 1984, 2001; Brauer and Castillo-
Chavez, 2001; Nedorezov and Utyupin, 2011 and many others). In situations when model (1) allows obtaining
good fitting it is possible to point out some other models which can give better results (Nedorezov, 2011,
2012). Attempts in modifying of Verhulst’ model (1) led to appearance of some other models. In particular,
within the framework of Gompertz’ model (Gompertz, 1825) it was assumed that influence of self-regulative
mechanisms is proportional to product )ln(xx :
x
Kx
dt
dxln . (2)
In model (2) both parameters are positive. If initial value Kx 00 then Ktx )( at t . If Kx 0
then Ktx )( . Note, expression )ln(xx describes influence of self-regulative mechanisms if and only if
1x (Nedorezov, 1997; Nedorezov and Utyupin, 2011). Model (2) can be modified with saving all basic
properties:
1
1ln
x
Kx
dt
dx. (3)
In model (2) all parameters are positive, 0,, constK . Below model (3) will be called as “theta-
Gompertz model”. Within the framework of model (3) influence of self-regulative mechanisms is described by
the expression )1(ln xx , and this expression was used for fitting of datasets.
In Svirezhev’ model (Svirezhev, 1987) negative influence of self-regulative mechanisms was described
with expression 3x , and increase of population size was proportional to 2x , 0, const :
32 xxdt
dx . (4)
Within the framework of theta-logistic model (Rosenzweig, 1969; Gilpin, Ayala, 1973) which is modification
of Verhulst’ model (1), respective expression has the form x , where is positive parameter,
1 const . In literature (see overview Nedorezov and Utyupin, 2011) it is possible to find a lot of various
modifications of pointed out models (1)-(4) but in most cases influence of self-regulative mechanisms is
described as monotonic increasing function with respect to population size in any power.
Use of physical ideas for modeling of ecological processes can be very useful. In various situations it
allows obtaining important results. On the other hand, as it was pointed out above, interaction between
biological individuals doesn’t look like colliding of absolutely elastic balls. There exists a lot of various types
of interaction between individuals: it can be a competition for food and space; it can be transmission of
diseases from one individual to another one etc. Moreover, scale of population size changing may be a non-
homogenous set: for biggest part of analyzed species Allee effect is observed (Allee, 1931; Odum, 1983).
107
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
Influence of this effect (existence of favorable levels of local population size) leads to changing of distribution
of individuals in habitat, and respectively to changing of a number of interactions (like average of interactions
in space) between individuals. Thus, these remarks allow concluding that question about types of functions
which can be applied for description of influence of self-regulative mechanisms on population dynamics is
open.
Problem pointed out above cannot be solved analyzing empirical or experimental datasets: self-regulation
contains a lot of various biological mechanisms, real population density is unknown amount and out of control
etc. Limits of favorable zone (Allee effect) are unknown too. Thus, this problem can be solved using
mathematical model of migrations only. In such a situation all basic population parameters are under the
control, and computer experiments can be provided with important artificial assumptions. One of such models
is described and analyzed below.
2 Model
2.1 Description
Let N be a total population size, and constN during the time of providing of computer experiments. Let 2nmZ be an integer rectangular lattice on the plane 2R :
}1,1:),{(2 mjnijiZ nm .
We’ll assume that local population size is determined in knots ),( ji of the lattice 2nmZ only. Denote it as
)(txij for 2),( nmZji at time moment t . Thus, for all time moments t , ...2,1,0t , the following relation
is truthful:
Ntxn
i
m
jij
1 1
)( .
It means that there are no migrations outside the domain 2nmZ ; birth and death processes are absent too. We’ll
say that two elements of the lattice ),( 11 ji , 222 ),( nmZji are neighboring knots if and only if the following
relation is truthful:
12121 jjii .
Within the framework of model it will be assumed that migration processes from the knot ),( ji can be
observed to neighboring knots only. Within the framework of considering model we’ll assume that every
individual with equal probabilities can migrate to nearest knots or stay in initial knot. Thus, 2.0p .
2.2 Initial conditions
As it was pointed out above, for modeling of migration processes it was assumed that total population size N
is constant; thus, theoretical population density was known and equal to nmN / . Initial population
state was modeled with discrete uniform distribution: every individual with equal probabilities could appear in
every knot of the lattice 2nmZ . After determination of initial positions the process of individual’s migrations
was started. During T time steps (for providing calculations it was assumed that 20000T ) model was run
free. It is important moment because we have to have on the lattice the situation which is determined by the
rules of population migration only, and doesn’t depend on the initial state of population.
2.3 Number of interactions between individuals
Let’s assume that at any fixed time moment t local population size ltxij )( . The basic question is: how can
we calculate number of interactions between individuals? First of all, it is naturally to assume that there are no
108
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
interactions between individuals from different knots. The second, if epizootics play most important role in
self-regulation, we have a good background for assumption that every individual contacts with all other
individuals in determined knot. Thus, in this case the number of paired interactions is equal to 2/)1( ll .
Below results of computer experiments for this assumption about number of interactions are called “first
dataset”. But it isn’t a unique type of grouping of individuals and their paired interactions (Odum, 1983;
Maynard, 1968, 1974).
Together with pointed out variant of local interactions of individuals we’ll consider the following
situation. It will be assumed that in every knot individuals can stay separately (i.e. without contacts with other
individuals in a knot), or can stay in pair, or form a group of three individuals. Let and be stochastic
variables with geometric distribution with parameter q . Number of pairs assumed to be equal to
}2/,min{* l . Number of groups with three individuals was equal to },3/)2{( ** l . Other
individuals ( ** 32 l ) were assumed to stayed separately. In this case the number of paired interactions
was determined as ** 3 . Below results of computer experiments for this assumption about number of
interactions are called “second dataset”.
3 Results of Calculations
After 20000 free steps of model during 20000 steps number of interactions between individuals was calculated
as average in space and time (for both variants). For every fixed time moment number of interactions was
calculated for every knot of lattice, and total sum of interactions was divided on product mn . All 20000
values of averages were summarized and divided on 20000 respectively. This procedure was repeated a certain
number of times for various values of population size.
Population size N was changed from zero up to 100000 with step 1000. Respectively, population
density ]10,0[ and was changed with step 0.1. Results of calculations of numbers of interactions between
individuals are presented on Fig. 1.
Fig. 1 Results of computer experiments: changing of numbers of interaction between individuals in two different cases with respect to changing of population density.
For fitting of obtained samples (Fig. 1) four different functions pointed out above were used. Deviations
between theoretical functions and obtained samples were tested on Normality and symmetry of distributions
(Kolmogorov – Smirnov test, Lilliefors test, Shapiro – Wilk test, Mann – Whitney test, and Wald – Wolfowitz
109
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
test), and on existence/absence of serial correlation (Draper and Smith, 1986, 1987; Lilliefors, 1967; Shapiro et
al., 1968; Bolshev and Smirnov, 1983; Hollander and Wolfe, 1973; Bard, 1974). Note, that all computer
experiments were provided independently; thus, if any curve gives good fitting of obtained datasets no serial
correlations must be observed.
Table 1 Results of testing on normality for deviations (first dataset).
Models: Parameters minQ * KS1 L2 SW3
Verhulst 4993.0 18.0 15.0p 01.0p 510p
Theta-Gompertz
636.0 34.2
7.06 2.0p 01.0p 00006.0p
Svirezhev 058.0 1420.4 05.0p 01.0p 510p
Theta-logistic 5.0 ,
0.2
0.18 15.0p 01.0p 510p
1KS – Kolmogorov – Smirnov test; 2L – Lilliefors test; 3SW – Shapiro – Wilk test; minQ * is minimal value of minimized functional form.
For the case when 100 nm , and parameter of geometric distribution q is equal to 2.0 , results of
testing on Normality of deviations for four classic models (more precisely, deviations between computer
results of calculation of number of interactions and functions in classic models which describe the influence of
self-regulative mechanisms) are presented in Tables 1 and 2. Parameters of functions were determined with
Least Square Method.
Results presented in table 1 show that best approximations were obtained with Verhulst and Theta-
logistic models. For both models 999992.02 R . For Svirezhev model 9379.02 R , and for Theta-
Gompertz model 9997.02 R . As we can see in all cases correlation coefficient 2R is very close to one,
and it means that we have rather good approximation for first dataset. On the other hand, Lilliefors test and
Shapiro – Wilk test showed that in all four considering cases with 1% significance level we have to reject
hypotheses about Normality of residuals. Thus, from the standpoint of traditional imagination about good
model (Bard, 1974) all functions are not suitable for fitting of first dataset.
Table 2 Results of testing on Normality for deviations (second dataset).
Models: Parameters minQ * KS1 L2 SW3
Verhulst 0642.0 268.38 05.0p 01.0p 510p
Theta-Gompertz
3947.0 , 3468.0
0.1957 15.0p 01.0p 510p
Svirezhev 0072.0 117.35 05.0p 01.0p 510p
Theta-logistic 3758.0 ,
1543.1
0.327 15.0p 01.0p 510p
1KS – Kolmogorov – Smirnov test; 2L – Lilliefors test; 3SW – Shapiro – Wilk test; minQ * is minimal value of minimized functional form.
110
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
The similar situation is observed for results presented in Table 2: Lilliefors test and Shapiro – Wilk test
showed that in all four considering cases with 1% significance level we have to reject hypotheses about
Normality of residuals. The best result was obtained for Theta-Gompertz model with 999258.02 R . For
Theta-logistic model 99876.02 R . For Svirezhev model this characteristics is rather small: 5548.02 R .
Like in previous case, from the standpoint of traditional imagination about good model (Bard, 1974) all
functions are not suitable for fitting of second dataset.
It is important to note that assumption about Normality of deviations between theoretical curves and
experimental datasets (in considering situation we have to talk about results of computer experiments) is rather
strong. Softer assumption is following: distribution density must be symmetric with respect to origin. Results
of checking of hypotheses about symmetry for both datasets are presented in Tables 3 and 4.
Table 3 Results of testing on symmetry for deviations (first dataset).
Models: KS1 WW2 MW3 Verhulst 1.0p 9968.0p 3573.0p
Theta-Gompertz 005.0p 3792.0p 016.0p
Svirezhev 1.0p 3196.0p 5335.0p
Theta-logistic 1.0p 5352.0p 5156.0p 1KS – Kolmogorov – Smirnov test; 2WW – Wald – Wolfowitz test; 3MW – Mann – Whitney test
In creation of conclusions about properties of datasets we’ll follow to the next basic principle: if one of
using tests gives a negative result we have to reject Null hypothesis, and it doesn’t depend on results obtained
with other tests. In particular, Kolmogorov – Smirnov test showed that we have to reject hypothesis about
symmetry of residuals obtained for Theta-Gompertz model with very small significance level (Table 3). In
other cases we cannot reject Null hypothesis about symmetry even with 10% significance level.
Table 4 Results of testing on symmetry for deviations (second dataset).
Models: KS1 WW2 MW3 Verhulst 1.0p 5586.0p 7887.0p
Theta-Gompertz 001.0p 009.0p 0009.0p
Svirezhev 1.0p 8422.0p 9423.0p
Theta-logistic 001.0p 0049.0p 0005.0p 1KS – Kolmogorov – Smirnov test; 2WW – Wald – Wolfowitz test; 3MW – Mann – Whitney test
Results presented in Table 4 allow concluding that deviations obtained for Theta-Gompertz model and
Theta-logistic model haven’t symmetric distributions: we have to reject hypotheses about symmetry even with
1% significance level. It is interesting to note that biggest values of probabilities were obtained for Svirezhev
model which has biggest value of minimizing functional form (Table 2).
As it was pointed out above for every value of population size (density ) computer experiments were
provided independently (Fig. 1). Additionally, we can consider population density as independent variable, as
111
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
a sequence of fixed time moments. Independence of computer experiments means that deviations between
theoretical and experimental results are independent stochastic variables. Thus, we cannot have correlation in
sequence of residuals if used model gives good fitting of dataset.
Critical values for Durbin – Watson test for 100 experimental points and one predictor variable are
following: 65.1Ld and 69.1Ud for 5% significance level and 52.1Ld and 56.1Ud for 1%
significance level (Draper and Smith, 1986, 1987). For first dataset we have the following results: for Verhulst
model 0268.0d ; for Svirezhev model 3901.0d ; for Theta-logistic model 0268.0d . Thus, in all
cases we have to reject hypothesis about absence of correlation with 1% significance level. For second dataset
we have the following results: for Verhulst model 0047.0d ; for Svirezhev model 0153.0d . For this
dataset we have also to reject hypothesis about absence of serial correlation.
For checking hypothesis about absence/existence of serial correlation we also used serial test (Draper and
Smith, 1986, 1987). For first dataset we have the following results: for Verhulst model number of positive
deviations is equal to 49, 491 n , number of negative deviation is equal to 51, 512 n , number of groups is
equal to 50, 50u , and 0965.0z (standard normal stochastic variable). Taking into account that
47.0}1.0{ zP we can conclude that observed combination of deviations with different signs and their
groups has very big probability. Thus, in this case we have no reasons for rejecting hypothesis about absence
of serial correlation. The same results we have for Theta-logistic model. For Svirezhev model 851 n ,
152 n , 2u , 554.9z ; thus, for this model combination of deviations with different signs and their
groups has very small probability, thus, we have to reject hypothesis about absence of serial correlation.
For second dataset we have the following results: for Verhulst model number of positive deviations is
equal to 81, 811 n , number of negative deviation is equal to 19, 192 n , number of groups is equal to 2,
2u , and 623.9z ; probability that z less or equal to -9.623 is very small, 002.0}3{ zP . For
Svirezhev model 851 n , 152 n , 2u , 554.9z . For both models we have to reject hypotheses
about absence of serial correlations in sequences of residuals.
4 Conclusion
Computer experiments with stochastic model of migrations of individuals on a plane under conditions that
population size is constant (no birth and death rates, no migrations out of and in to considering domain,
homogenous structure of locations) allowed obtain two various datasets of interactions between individuals.
First dataset was obtained for the case when in every location every individual connected with all other
individuals. Second dataset was obtained for the situation when in locations individuals could stay separately
or organize group in two or three individuals.
A lot of classic models of population dynamics were constructed under the assumption that influence of
self-regulative mechanisms is determined by numbers of interactions between individuals. Approximation of
obtained datasets by various functions describing influence of self-regulative mechanisms (in Verhulst, Theta-
Gompertz, Svirezhev, and Theta-logistic models) showed that all functions are not suitable for fitting of
second dataset. For the first dataset Verhulst model and Theta-logistic model can be used for fitting. More
precisely, last models have good backgrounds for it; but from the standpoint of traditional imagination about
good and bad models (Bard, 1974) Verhulst and Theta-logistic equations are not suitable for approximation.
When requirements for used model are not so strong (in particular, when distribution of residuals must be
symmetric only, and in sequence of residuals serial correlation cannot be observed) these model can be used
for fitting.
Obtained results don’t allow concluding that used models cannot be applied for modeling of population
dynamics. We obtained the background for conclusion that within the frameworks of considered models
112
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 106-113
IAEES www.iaees.org
influence of intra-population self-regulative mechanisms haven’t strong correlation with numbers of
interactions between individuals.
References
Allee WC. 1931. Animal Aggregations: A Study in General Sociology. Chicago University Press, Chicago,
USA
Bard Y. 1974. Nonlinear Parameter Estimation. Academic Press, San Francisco, London, USA
Bolshev LN, Smirnov NV. 1983. Tables of Mathematical Statistics. Nauka, Moscow, Russia
Brauer F, Castillo-Chavez C. 2001. Mathematical Models in Population Biology and Epidemiology. Springer-
Verlag, NY, USA
Draper NR, Smith H. 1986. Applied Regression Analysis. V.1. Finance and Statistics, Moscow, Russia.
Draper NR, Smith H. 1987. Applied Regression Analysis. V.2. Finance and Statistics, Moscow, Russia
Gause GF. 1934. The Struggle for Existence. Williams and Wilkins, Baltimore, USA
Gilpin ME, Ayala FJ. 1973. Global models of growth and competition. Proceedings of the National Academy
of Sciences USA, 70: 3590-3593
Gompertz B. 1825. On the nature of the function expressive of the law of human mortality and on a new model
of determining life contingencies. Philosophical Transactions of the Royal Society London, 115: 513-585
Hollander M, Wolfe DA. 1973. Nonparametric statistical methods. John Wiley & Sons, New York-Sydney-
Tokyo-Mexico City, USA
Isaev AS, Khlebopros RG, Nedorezov LV, et al. 1984. Forest Insect Population Dynamics. Nauka,
Novosibirsk, Russia
Isaev AS, Khlebopros RG, Nedorezov LV, et al. 2001. Population Dynamics of Forest Insects. Nauka,
Moscow, Russia
Lilliefors HW. 1967. On the Kolmogorov-Smirnov test for normality with mean and variance
unknown. Journal of the American Statistical Association 64: 399-402
Maynard SJ. 1968. Mathematical Ideas in Biology. Cambridge University Press, Cambridge, USA
Maynard SJ. 1974. Models in Ecology. Cambridge University Press, Cambridge, USA
Nedorezov LV. 1997. Course of Lectures on Ecological Modeling. Siberian Chronograph, Novosibirsk, Russia
Nedorezov LV. 2011. Analysis of some experimental time series by Gause: Application of simple
mathematical models. Computational Ecology and Software, 1(1): 25-36
Nedorezov LV. 2012. Gause’ Experiments vs. Mathematical Models. Population Dynamics: Analysis,
Modelling, Forecast, 1(1): 47-58
Nedorezov LV, Utyupin YuV. 2011. Continuous-Discrete Models of Population Dynamics: An Analytical
Overview. State Public Scientific-Technical Library of Russian Academy of Sciences, Novosibirsk, Russia
Odum EP. 1983. Basic Ecology. Saunders College Pub., Philadelphia, USA
Pielou EC. 1977. Mathematical Ecology. John Wiley and Sons, NY, USA
Rosenzweig ML. 1969. Why the prey curve has a hump. American Naturalist, 103: 81-87
Shapiro SS, Wilk MB, Chen HJ. 1968. A comparative study of various tests of normality. Journal of the
American Statistical Association 63: 1343-1372
Svirezhev YuM. 1987. Nonlinear Waves, Dissipative Structures and Catastrophes in Ecology. Nauka, Moscow,
Russia
Verhulst PF. 1838. Notice sur la loi que la population suit dans son accroissement. Corresp. Math. et Phys., 10:
113-121
113
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
Article
Multivariate statistical analysis of surface water chemistry: A case
study of Gharasoo River, Iran
MH Sayadi 1, A Rezaei1, MR Rezaei1, K Nourozi2 1Environmental Department, University of Birjand, Birjand, Iran 2Department of Environmental Protection Kermanshah, Iran
E-mail: [email protected]
Received 8 May 2014; Accepted 15 June 2014; Published online 1 September 2014
Abstract
Regional water quality is a hot spot in the environmental sciences for inconsistency of pollutants. In this paper,
the surface water quality of the Gharasoo River in western Iran is assessed incorporating multivariate statistical
techniques. Parameters like EC, TDS, pH, HCO3-, Cl-, SO4
2-, Ca2+, Mg2+ and Na+ were analyzed. Principal
component and factor analysis is showed the parameters generated 3 significant factors, which explained
73.06℅ of the variance in data sets. Factor 1 may be derived from agricultural activities and subsequent release
of EC, TDS, SO42- and Na+ to the water. Factor 2 could be influenced by domestic pollution and explained the
deliverance of HCO3-, Cl- and Mg2+ into the water. Factor 3 contains hydro-geochemical variable Ca2+ and pH,
originating from mineralization of the geological components of bed sediments and soils of watershed area.
Likewise, the clustering analysis generated 3 groups of the stations as the groups had similar characteristic
features. Pearson correlation analysis showed significant correlations between HCO3- and Mg2+ (0.775), Ca2+
(0.552) as well as TDS and Na+ (0.726). With reference to multivariate statistical analyses it can be concluded
that the agricultural, domestic and hydro-geochemical sources are releasing the pollutants into the Gharasoo
River water.
Keywords anthropogenic activities; geological components; Gharasoo River; PCA; water quality.
1 Introduction
1 Introduction
The surface water quality is truly a sensitive issue today because of its effects on human health and aquatic
ecosystems. Rivers are highly vulnerable to pollution attributing to their role in carrying off the municipal and
industrial wastewater and runoff from agriculture in their vast drainage basins. Anthropogenic influences, as
well as natural processes, deteriorate surface water and impair their use for drinking, industrial, agricultural
Proceedings of the International Academy of Ecology and Environmental Sciences ISSN 22208860 URL: http://www.iaees.org/publications/journals/piaees/onlineversion.asp RSS: http://www.iaees.org/publications/journals/piaees/rss.xml Email: [email protected] EditorinChief: WenJun Zhang Publisher: International Academy of Ecology and Environmental Sciences
IAEES
and recre
Rezaei, 2
The
on receiv
water is
weatherin
From
widely u
et al., 20
to invest
natural in
similariti
sources (
possible
2 Materi
2.1 Stud
The Gha
is approx
diplomat
collects w
lies betw
above se
Rainfall
Proceedings
eational purpo
2014).
e environmen
ving water ar
most vulne
ng of crystal
m many met
used, being ca
06; Wenning
tigate the Gh
nfluences on
ies or dissim
(natural and
sources deter
ial and Meth
y Area
rasoo River s
ximately 20.7
tic area in we
water from K
ween latitudes
ea level (Fig.
occurs during
of the Internati
oses (Sayadi
ntal impacts o
re numerous
erable to po
materials, an
thods availab
apable of dete
g and Erickso
harasoo river
the river wat
milarities betw
anthropogeni
rmining wate
hods
system forms
7 kilometers a
estern Iran. G
Kermanshah a
s 46° 36´ - 47
. 1). The me
g the autumn
Fig
ional Academy
et al., 2010;
of agricultural
and inputs o
llution due
nd anthropoge
ble, Cluster A
ecting similar
on, 1994). In t
r water quali
ter quality. Th
ween monitor
ic) on water
er quality para
with joining
and runs thro
Gharasoo Rive
and Kurdista
7° 37´ N and
ean annual te
and winter (D
g. 1 Location o
y of Ecology and
Sayadi and S
l runoff, mun
of contamina
to natural p
enic influence
Analysis (CA
rities among s
this work, PC
ity and discr
he objective
ring periods a
quality param
ameters of the
of two main
ough the city
er is one of th
an provinces
d longitudes 3
emperature an
December to
f sample sites o
d Environmenta
Sayyed, 2011;
nicipal wastew
ants can be h
processes, su
es (Sundaray
A) and Princ
samples and/o
CA and hierar
riminate relat
of the study
and monitori
meters, and (
e Gharasoo R
branches of M
of Kermansh
he most impo
and delivers
34° 00´- 34°
nd rainfall ar
March).
on the Gharasoo
al Sciences, 20
; Sayyed and
water and ind
hazardous to
uch as precip
et al., 2006).
cipal Compon
or variables (
rchical cluste
tive magnitu
is to extract i
ng sites, (2)
3) source ide
River.
Merek and R
hah. Kermans
ortant tributar
it to Saymar
91´ E with t
re 14°C and
o River.
14, 4(3): 114-12
w
Sayadi, 2011
dustrial efflue
the environm
pitation inpu
nent Analysi
(Acque et al.,
er analysis ha
ude of anthro
information a
the influence
entification fo
Raz Avar Rive
shah city is l
ries of Sayma
reh River. Th
the height of
456.8 mm,
22
www.iaees.org
1; Sayadi and
ent discharge
ment. Surface
uts, erosions,
is (PCA) are
1995; Wang
ave been used
opogenic and
about; (1) the
e of possible
or estimating
ers. Its length
ocated in the
areh River. It
he study area
f 1322 meters
respectively.
d
e
e
,
e
g
d
d
e
e
g
h
e
t
a
s
.
115
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
Data sets of 9 parameters of the water quality were monitored monthly over a period of 2009-2010.
Monitoring stations are shown in (Fig. 1). The selected parameter for the determination of water quality
characteristics were EC, PH, TDS, bicarbonate (HCO3-), chloride (Cl‾), sulfate (SO4
2-), calcium (Ca2+),
magnesium (Mg2+) and sodium (Na+). The parameters were analyzed according to standard methods (APHA-
AWWA-WPCF, 1985; APHA, 1999). The results were evaluated via multivariate statistical analysis
techniques. All statistical computations were made using SPSS statistical software.
2.2 Principal Component Analysis (PCA)
PCA is designed to transform the original variables into new and uncorrelated variables called the principal
components, which are linear combinations of the original variables (Zhang, 2011; Vieira, 2012). It provides
information on the most significant parameters due to spatial and temporal variations that describes the whole
data set by excluding the less significant parameters with minimum loss of the original information (Helena et
al., 2000; Kannel et al., 2007). The principal component can be expressed as
Zij=ai1 x1j + ai2 x2j + … + aim xmj (1)
where z is the component score, a is the component loading, x is the measured value of the variable, I is the
component number, j is the sample number, and m is the total number of variables.
Factor analysis follows principal component analysis. The main purpose of factor analysis is to reduce the
contribution of less significant variables and to simplify even more the data structure coming from the
principal component analysis. This purpose can be achieved by rotating the axis defined by principal
component analysis according to well established rules, and constructing new variables, also called vary
factors. A small number of factors will usually account for approximately the same amount of information as
does the much larger set of original observations (Shrestha and Kazama, 2007). The Factor analysis can be
expressed as:
Zji = af1 f1i + af2 f2i + af3 f3i + … + afmfmi + efi (2)
where z is the measured value of a variable, a is the factor loading, f is the factor score, e is the residual term
accounting for errors or other sources of variable number, and m is the total number of factors.
2.3 Cluster Analysis (CA)
CA is a multivariate technique, whose primary purpose is to classify the objects of the system into categories
or clusters based on their similarities (Zhang, 2012), and the objective is to find an optimal grouping for which
the observation or objects within each cluster are similar, but the cluster is dissimilar to each other.
Hierarchical clustering is the most common approach in which clusters are formed sequentially. The most
similar objects are first grouped, and these initial groups are merged according to their similarities. Eventually
as the similarity decreases all subgroups are merged into a single cluster. CA was applied to surface water
quality data using a single linkage method. In the single linkage method, the distances or similarities between
two clusters A and B are defined as the minimum distance between a point A and a point in B:
D (A, B) = min {d (xi+xj), for xi in A and xj in B} (3)
where d (xi +xj) is the Euclidean distance in (3). At each step the distance is found in every pair of clusters and
the two clusters with smallest distance are merged. When over two clusters are merged the procedure is
repeated for the next step: the distances between all pairs of clusters are calculated again, and the pair with the
minimum distance is merged into a single cluster. The result of a hierarchical clustering procedure can be
displayed graphically using a tree diagram, also known as a dendrogram, which shows all the steps in the
hierarchical procedure (Alkarkhi et al., 2008; Johnson and Wichern, 2002).
116
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
3 Results and Discussion
Table 1 summarizes briefly the mean, maximum and minimum values besides standard deviation and variance
of the 9 measured parameters in the river water samples from the five stations. It is interesting to note that the
high standard deviations of the parameters indicating changeability in chemical composition between the
samples, shows the temporal variations which appears by lithogenic and anthropogenic sources.
Table 1 Simple statistical analysis of water quality parameters at different locations on the Gharasoo River.
Station EC pH TDS HCO3- Cl SO4
2- Ca Mg Na Station1
Mean Std. Variance Minimum Maximum
372 53.8 290 172 437
7.8 0.46 0.21 6.53 8.57
2.4 35.2 1.24 108 280
3.3 0.55 0.31 2.31 5.06
0.520.220.050.161.10
0.49 0.27 0.07 0.1 1.29
2.7 0.42 0.18 1.91 3.41
1.3 0.38 0.15 0.56 2.24
0.36 0.17 0.03 0.09 0.91
Station2
Mean Std. Variance Minimum Maximum
437 96 91.70 329 661
7.7 0.46 0.22 6.70 8.52
280 61 37.45211 423
3.73 0.8 0.6 2.56 5.43
0.560.250.060.160.96
0.49 0.27 0.07 0.14 0.92
2.9 0.43 0.19 2.01 3.55
1.47 0.65 0.42 0.78 2.80
0.43 0.31 0.09 0.16 1.16
Station3 Mean Std. Variance Minimum Maximum
404 56.61 321 320 540
7.79 0.36 0.13 7.04 8.40
285 36.18130 205 346
3.43 0.68 0.46 2.30 5.11
0.660.140.020.380.91
0.59 0.28 0.0810.20 1.36
2.94 0.42 0.18 2.37 3.37
1.42 0.42 0.18 0.81 2.40
0.38 0.08 0.01 0.25 0.59
Station4 Mean Std. Variance Minimum Maximum
434 111 124 312 663
7.86 0.37 0.13 7.19 8.66
275 71.5151.14199 424
3.58 0.93 0.87 2.10 6.46
0.520.220.050.221.00
0.72 0.52 0.28 0.16 2.71
2.77 0.50 0.25 1.41 3.49
1.56 0.58 0.34 0.56 3.00
0.55 0.53 0.28 0.06 2.20
Station5 Mean Std. Variance Minimum Maximum
494 491 241 340 520
7.37 0.13 0.02 7.32 7.80
336 199 397 320 390
4.00 0.29 0.08 3.44 4.38
0.430.040.010.360.50
0.91 0.14 0.02 0.64 1.11
3.00 0.08 0.01 2.80 3.13
1.94 0.41 0.17 1.00 2.50
0.45 0.03 0.00 0.39 0.49
3.1 Application of PCA to Gharasoo River data set
A particular problem in the surface water quality monitoring is the complexity associated with analyzing a
large number of measured variables (Saffran et al., 2001). Therefore, in this study, surface water quality data
were grouped using FA. The correlation matrix of variables was generated and factors were extracted by the
centroid method, rotated by Varimax. From the results of the FA, the first three eigenvalues were found to be
bigger than 1 (Fig. 2). According to the Fig. 2 and a subsequent interpretation of the factor loadings, the first
three components were extracted and the other components have been eliminated.
117
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
Fig. 2 Screen plot of the eigenvalue and component number.
Table 2 presents the total variance explained by the first three factors for both related and unrelated factor
loadings. The parameter loading three factors in the two from FA associated with each factor stations are well
defined and contribute slightly to other factors, which help not only in the interpretation of the results but also
in the identification of anthropogenic sources of pollution from the surface water quality data. FA generated
three significant factors, which explained 73.06℅ of the variance in data sets, where a correlation greater than
0.75 is considered “strong”; 0.75-0.50, “moderate”; and 0.50-0.30, as “weak” significant factor loading (Liu et
al., 2003).
Table 2 Extracted values of various FA parameters.
Component Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings
Total % of Variance Cumulative % Total % of Variance Cumulative %
1 3.85 42.86 42.86 2.81 31.23 31.23
2 1.53 17.03 59.89 2.47 27.49 58.72
3 1.18 13.16 73.06 1.29 14.33 73.06
Table 3 Loadings of 9 experimental variables on 3 significant Principal components,
rotated factor loadings matrix
Variables Factor 1 Factor 2 Factor 3 EC 0.831 0.360 0.087 pH -0.180 0.180 -0.687 TDS 0.858 0.306 0.102 HCO3
- 0.102 0.844 0.422 Cl 0.220 0.720 -0.221 SO4
2- 0.740 -0.184 0.198 Ca -0.040 0.389 0.716 Mg 0.240 0.820 0.046 Na 0.829 0.361 -0.136
118
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
As shown in tables 2 and 3, the first factor (Factor 1), accounted for 31.23 of the total variance, was high
positive loading in EC, TDS, SO42-and Na+ which were 0.831, 0.858, 0.740 and 0.829 respectively. This factor
represents the contribution of nonpoint pollution from agricultural areas. In these areas, farmers use sulfate
fertilizers, and the stream receives sulphate via surface runoff and irrigation waters. As a result shown increase
in SO42- concentrations may be due to agricultural activities (Krouse, 1997). The contribution of Na+ to this
factor can be considered a result of action-exchange processes in soil–water interface (Guo and Wang, 2004).
Factor 2 explains 27.49% of the total variance and is the positively correlated with HCO-, Cl¯ and Mg2+. This
factor represents the contribution of point pollution and the physico-chemistry of the stream. While point
pollution is from domestic wastewater, nonpoint pollution is from agricultural and livestock farms. Mg2+ is a
basic metal which increases alkalinity of the environment (Razmkhah et al., 2010). This factor may also be due
to anthropogenic activities such as domestic waste water or influents. Nevertheless, the release of domestic
effluents into the river water caused the dramatic Cl- increase. The loading for factor 3 was 18.92% with Ca2+
and pH. Thus, this factor contains hydro-geochemical variable Ca2+, originating, at a first glance, from
mineralization of the geological components of soils as well as moderate decrease of pH concentration. The
contribution of Ca2+ to this factor can be considered a result of action-exchange processes in soil–water
interface (Guo and Wang, 2004) as the results demonstrated an increase in EC, TDS, SO42- and Cl¯
concentrations due to agricultural and domestic waste water activities. Sources of dissolved SO42- in natural
river waters may include dissolution of sedimentary sulfates, oxidation of both sulfide minerals and organic
materials, and anthropogenic inputs.
3.2 Pearson correlation
Statistical analysis using Pearson correlation showed that the parameters in the water samples collected from
Gharasoo river were weak and moderately correlated to each other at p <0.01 and p <0.05 levels. A significant
positive correlation was found to exist between EC and TDS (0.918), HCO3- (0.432), Cl¯ (0.381), SO4
-2 (0.411),
Mg2+ (0.421), and a positive correlation was found between EC and Na+ (0.721) at p <0.01 (Table 4).
Similarly, there were significant correlations between TDS and HCO3- (0.387), Cl¯ (0.369), SO4
2-(0.434),
Mg2+ (0.385) and a positive correlation between TDS and Na+ (0.726) at p <0.01 in the collected water
samples of the study region. The level of TDS reflects the pollutant burden of the water. High levels of
dissolved and suspended solids in water systems increase the biological and chemical oxygen demand
(Jonnalagadda and Mhere, 2001). Similarly, some correlations were also observed (Table 4) between HCO3-
and Cl- (0.373), Na+ (0.387) and a positive strong correlation was found between HCO3- and Mg2+ (0.775)
and Ca2+ (0.552) at p <0.01. Likewise, Li and Zhang (2009) indicated a strong positive correlation between
HCO3-, Ca2+ and Mg2+in Geochemistry of the upper Han River basin, China. There were as well significant
positive correlations between Cl¯- Mg2+ (0.506), Cl--Na+ (0.453), SO4-2 - Na+ (0.529), and Mg2+ - Na+ (0.459)
at p <0.01 level (Table 4). Chloride concentration is higher in wastewater than raw water because sodium
chloride, the commonest component of the human diet passes unchanged through the digestive system (WHO,
2008).
It is interesting to note that in this study there is no significant correlation between pH and other
parameters. Similarly, Chigor et al. (2012) exhibited nil correlation between pH and other contaminated
parameters in surface water sources used for drinking and irrigation in Zaria, Nigeria.
3.3 Hierarchical cluster analysis (HCA)
Spatial similarity and monitoring stations grouping is shown in Fig. 3. In this study, the classification of
monitoring stations was performed incorporating HCA, and a dendrogram was composed.
119
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
Table 4 Pearson correlation between different water quality parameters of the study site.
Discriminate variables
EC pH TDS HCO3- Cl SO4
-2 Ca Mg Na
EC 1.00
pH -0.126 1.00 TDS 0.918** -0.177 1.00 HCO3
- 0.432** -0.140 0.387** 1.00 Cl 0.381** 0.070 0.369** 0.373** 1.00 SO4
-2 0.411** -0.086 0.434** -0.039 0.014 1.00 Ca 0.169 -0.096 0.148 0.552** 0.175 0.204* 1.00 Mg 0.421** -0.017 0.385** 0.775** 0.506** 0.166 0.180* 1.00 Na 0.721** -0.017 0.726** 0.387** 0.453** 0.529** 0.000 0.459** 1.00
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).
Fig. 3 Dendrogram of the CA according to single linkage method.
The clustering procedure generated 3 groups of stations in a very convincing way, as the sites in these
groups have similar characteristic features and natural background source types. Cluster 1 (Stations 2, 3 and 4),
Cluster 2 (Station 1) and Cluster 3 (Station 5) correspond to a relatively low to high polluted regions. Hence,
the temporal variation in the Gharasoo river water quality was greatly determined by agricultural and
municipal activities as well as lithogenic sources which confirm the result of the PCA. In fact, Fig. 3 shows
that the patterns of pollution sources of Gharasoo river water.
4 Conclusion
In this study, multivariate statistical methods including factor, principal component and cluster analysis were
applied to surface water quality data sets obtained from the Gharasoo River in Iran. The results suggest that
anthropogenic activities such as agricultural and domestic pollution sources and lithogenic activities had
significant effects on water quality. Three factors explaining the 73.06℅ of the total variance in the surface
water quality data set were determined. Based on the above results, Factor 1 may be derived from agricultural
120
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
activities and release the EC, TDS, SO42- and Na+ to the environment. Factor 2 could be influenced by
domestic pollution and explained the deliverance of HCO-3, Cl¯ and Mg2+ into the surface water of Gharasoo
River. Factor 3 contains hydro-geochemical variable Ca2+ and pH, originating from mineralization of the
geological components of bed sediments and soils of watershed area. Cluster analysis grouped the monitoring
stations into 3 clusters of similarity based upon water quality characteristics at different stations. These results
reveal that agricultural, domestic and hydro-geochemical sources are responsible for pollutions in terms of
water quality in Gharasoo River.
Acknowledgment
The authors would like to appreciate the Department of Environment, Head Office, Kermanshah city for their
cooperation and support. Authors are appreciated the authorities of Research Council and Faculty of Natural
Resources and Environment, University of Birjand, due to their sincere cooperation. We also like to thank Dr.
Mrs. Mahavash F. Kavian for editing the paper.
References
Acque RS, Battegazzore M, Renoldi M. 1995. Integrated chemical and biological evaluation of the quality of
the River Lambro (Italy). Water Air and Soil Pollution, 83(3-4): 375-390
Alkarkhi A FM, Ahmad A, Easa A M. 2008. Assessment of surface water quality of selected estuaries of
Malaysia: multivariate statistical techniques. The Environmentalist, 29(3): 255-262
APHA-AWWA-WPCF. 1985. Standard methods for the Examination of Water and Wastewater (16th ed),
American Public Health Association, American Water Works Association, Water Pollution Control
Federation, USA
American Public Health Association (APHA). 1999. Standard methods for the examination of water and
wastewater (20th ed.). Washington D.C.: APHA, AWWA, Washington DC, USA
Chigor VN, Umoh VJ, Okuofu CA, et al. 2012. Water quality assessment: surface water sources used for
drinking and irrigation in Zaria, Nigeria are a public health hazard. Environmental monitoring and
assessment, 1845: 3389-3400
Guo H, WangY. 2004. Hydrogeochemical processes in shallow quaternary aquifers from the northern part of
the Datong Basin, China. Applied Geochemistry, 19(1): 19-27
Helena B, Pardo R, Vega M, Barrado E. 2000. Temporal evolution of groundwater composition in an alluvial
aquifer (Pisuerga River, Spain) by principal component analysis. Water Research, 34(3): 807-816
Johnson RA, Wichern DW. 2002. Applied Multivariate Statistical Analysis (Vol. 5). 1-13, Prentice Hall
Upper Saddle River, NJ, USA
Jonnalagadda SB, Mhere G. 2001. Water quality of the odzi river in the Eastern Highlands of Zimbabwe.
Water Research, 35(10): 2371-2376
Kannel PR, Lee S, Kanel SR, Khan S P. 2007. Chemometric application in classification and assessment of
monitoring locations of an urban river system. Analytica Chimica Acta, 582(2): 390-399
Krouse HR. 1997. Application of the stable isotope composition of S O 4 to tracing anomalous TDS in Nose
Creek , southern Alberta , Canada IRegional Bow River. Applied Geochemistry, 12(97): 567-575
Li S, Zhang Q. 2009. Geochemistry of the upper Han River basin, China. 2: Seasonal variations in major ion
compositions and contribution of precipitation chemistry to the dissolved load. Journal of hazardous
materials, 170(2-3): 605-611
121
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 114-122
IAEES www.iaees.org
Liu CW, Lin KH, Kuo YM. 2003. Application of factor analysis in the assessment of groundwater quality in a
blackfoot disease area in Taiwan. The Science of the Total Environment, 313(1-3): 77-89
Razmkhah H, Abrishamchi A, Torkian A. 2010. Evaluation of spatial and temporal variation in water quality
by pattern recognition techniques: A case study on Jajrood River (Tehran, Iran). Journal of environmental
management, 91(4): 852-860
Saffran K, Environment A, Cash K, Canada E. 2001. Canadian water quality guidelines for the protection of
aquatic life, CCME water quality Index 1, 0. User‘s Manual. Excerpt from Publication No. 1299, Canada
Sayadi MH, Rezaei MR. 2014. Impact of land use on the distribution of toxic metals in surface soils in Birjand
city, Iran. Proceedings of the International Academy of Ecology and Environmental Sciences, 4(1): 18-29
Sayadi MH, Sayyed MRG, et al. 2010. Short-term accumulative signatures’ of heavy metal in river bed
sediments, Tehran Iran. Environmental Monitoring and Assessment, 162: 465-473
Sayadi MH, Sayyed MRG. 2011. Comparative assessment of baseline concentration of the heavy metals in the
soils of Chitgar Industrial Area Tehran (Iran) with the comprisable reference data. Environmental Earth
Sciences, 63(6): 1179-1188
Sayyed MRG, Sayadi MH. 2011. Variations in the heavy metal accumulations within the surface soils from the
Chitgar Industrial Area of Tehran (Iran). Proceedings of the International Academy of Ecology and
Environmental Sciences, 1(1): 27-37
Shrestha S, Kazama F. 2007. Assessment of surface water quality using multivariate statistical techniques: A
case study of the Fuji river basin, Japan. Environmental Modelling & Software, 22(4): 464-475
Sundaray SK, Panda UC, Nayak BB, Bhatta D. 2006. Multivariate statistical techniques for the evaluation of
spatial and temporal variations in water quality of the Mahanadi river-estuarine system (India)--a case
study. Environmental geochemistry and health, 28(4): 317-330
Vieira VMNCS. 2012. Permutation tests to estimate significances on Principal Components Analysis.
Computational Ecology and Software, 2(2): 103-123
Wang YYS, Lou ZZP, Sun CC C, Wu MML, Han SH. 2006. Multivariate statistical analysis of water quality
and phytoplankton characteristics in Daya Bay, China, from 1999 to 2002. Oceanologia, 48(2)
Wenning RJ, Erickson GA. 1994. Interpretation and analysis of complex environmental data using
chemometric methods. TrAC Trends in Analytical Chemistry, 13(10): 446-457
World Health Organization. 2008. Guideline to drinking water quality (3rd ed., Vol. 1). World Health
Organization, Geneva, Switherlands
Zhang WJ. 2011. Simulation of arthropod abundance from plant composition. Computational Ecology and
Software, 1(1): 37-48
Zhang WJ. 2012. Computational Ecology: Graphs, Networks and Agent-based Modeling. World Scientific,
Singapore
122
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
Article
Detecting barriers and facilities to species dispersal: Introducing
sloping flow connectivity
Alessandro Ferrarini
Department of Evolutionary and Functional Biology, University of Parma, Via G. Saragat 4, I-43100 Parma, Italy
E-mail: [email protected], [email protected]
Received 28 April 2014; Accepted 2 June 2014; Published online 1 September 2014
Abstract
Connectivity in ecology deals with the problem of how biotic dispersals can happen, given actual landscape
properties and species presence/absence over such landscape. Recently I have introduced a modelling approach
(flow connectivity) to ecological connectivity that is alternative to circuit theory, and is able to fix the weak
point of the “from-to” connectivity approach. In addition, I’ve introduced “reverse flow connectivity” that
couples evolutionary algorithms to partial differential equations in order to fix the problem of subjectivity in
the attribution of friction values to landscape categories. I’ve also showed that flow connectivity can be used to
predict biotic movements happened in the past (backward flow connectivity). To date, there has been little
effort by conservation scientists towards detecting restoration opportunities by mapping barriers that strongly
reduce movement potential. In this paper, I introduce a new kind of theoretical and modelling approach called
“sloping flow connectivity”. The goal of such proposal is to individuate and map barriers and facilities to
species dispersals over the landscape. I define here a barrier as a landscape feature that impedes biotic
movements, the removal of which would increase the potential for biotic shifts. Using sloping flow
connectivity, it’s possible to plan greenways and ecological networks in an effective manner, since it is able to
enhance the real potential of each landscape elements to facilitate or obstruct both directional and overall
species movements.
Keywords biotic flows; dispersal facilities; flow connectivity; gene flow; landscape barriers; landscape
connectivity; partial differential equations; species dispersal.
1 Introduction
1 Introduction
Predicting how animals disperse is a pivotal issue for the management and conservation of fragmented
populations. Landscape heterogeneity and fragmentation affect how organisms are distributed in the landscape
(Fahrig and Merriam, 1985; Kennedy and Gray, 1997), determine the chance of a patch being colonized
(Hanski and Ovaskainen, 2000), reduce inbreeding in small populations and maintains evolutionary potential
Proceedings of the International Academy of Ecology and Environmental Sciences ISSN 22208860 URL: http://www.iaees.org/publications/journals/piaees/onlineversion.asp RSS: http://www.iaees.org/publications/journals/piaees/rss.xml Email: [email protected] EditorinChief: WenJun Zhang Publisher: International Academy of Ecology and Environmental Sciences
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
(Couvet, 2002). In order to predict dispersal, it is important to not only consider an organism’s dispersal
capabilities, but also the complex interactions between its behaviour and the landscape pattern and use.
The modelling of animal dispersal provides a useful tool for investigating these complex interactions, and
it is an essential goal for biotic conservation planning (Kareiva and Wennergren, 1995; King and With, 2002).
Due to the difficulty in gathering experimental results on species dispersal, simulation models have become a
cost-effective approach to predict dispersal dynamics (Tischendorf, 1997; Wiegand et al., 1999; Tischendorf
and Fahrig, 2000). Simulation models with spatially-explicit landscapes enable the integration of the
relationships between species and the landscape, and provide representation of the spatial elements that
promote or constrain dispersal. Several dispersal models with spatially explicit landscapes have been
developed. Some consider dispersal behaviour according to habitat affinity or physiological states in order to
predict animal movements and provide guidelines for landscape and wildlife management (Gustafson and
Gardner, 1996; With et al., 1997; Gardner and Gustafson, 2004).
Recently I have introduced a modelling approach (flow connectivity; Ferrarini, 2013a) to ecological
connectivity that is alternative to circuit theory (McRae, 2006; McRae and Beier, 2007; McRae et al., 2008),
and is able to fix the weak point of the “from-to” connectivity approach. Landscape connectivity as estimated
by circuit theory relies on a strong assumption that is possibly untrue, unproven or very challenging to be
demonstrated: species dispersals are “from-to” movements, i.e. from source points (patches) of the landscape
to sink ones. Source and sinks are suitable areas present within a matrix that is partially or completely hostile
to the species. There are two aspects of this approach that are questionable. First, a source-sink habitats model
can be suitable to describe lowland landscapes where few suitable patches (e.g. protected areas) are surrounded
by a dominant, hostile (or semi-hostile) anthropogenic landscape. By the way, can we think the same of
mountain and hilly landscapes? Such landscapes are not composed of source and sink habitats, instead they’re
a continuum with a natural matrix where the source-sink habitats model loses its rationale. Second, assuming
that a species aims to go from “patch A” to “patch B” means that such species is supposed to plan such
dispersal path (i.e. global optimization). This could be true for short-range dispersals where the final point is
visible from the starting one, but for wide-range movements, and for plant species in particular, the dispersal
model postulated by circuit theory is unsuitable.
In addition, I’ve introduced “reverse flow connectivity” (Ferrarini, 2014a) that couples evolutionary
algorithms to partial differential equations in order to fix the problem of subjectivity in the attribution of
friction values to landscape categories. I’ve also showed that flow connectivity can be used to predict biotic
movements happened in the past (backward flow connectivity; Ferrarini, 2014b).
In this paper, I introduce a new kind of theoretical and modelling approach called “sloping flow
connectivity”. The goal of such proposal is to individuate barriers (to be removed) and facilities (to be
conserved) to species dispersal over the landscape. The reason behind sloping flow connectivity is that it
makes possible to plan greenways and ecological networks in an effective manner, since it is able to enhance
the real potential of each landscape elements to facilitate or obstruct directional and overall species movements.
2 Sloping Flow Connectivity: Mathematical Formulation
Let ( , , , )L x y z t be a real 3D landscape at generic time t, where [1,..., ]L n . In other words, L is a generic
(categorical) landcover or land-use map with n classes. At time T0,
0 0( , , , )L L x y z t (1)
Let be the landscape friction (i.e. how much each land parcel is unfavourable) to the species under
study. In other words, ( )L is a function that associates a friction value to each pixel of L. At time T0,
( )L
124
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
0 0( )L (2)
Let ( , , ( ))sL x y L be a landscape where, for each pixel, the z-value is equal to the friction for the species
under study. In other words, Ls is a 3D fictional landscape with the same coordinates and geographic
projection as L, but with pixel-by-pixel friction values in place of real z-values. Higher elevations represents
areas with elevated friction to the species due to whatever reason (unsuitable landcover, human disturbance
etc), while lower altitudes represent the opposite.
True-to-life coefficients for landscape friction can be calculated as in Ferrarini (2014a), where I defined P
as the predicted path for the species over the fictional landscape Ls, and P* the real path followed by the
species as detected by GPS data-loggers or in situ observations. The bias B between P and P* is hence
calculated as
*mod( )B Pdx P dx (3)
where the function mod indicates the module of the difference.
Hence:
* *
* *
where >
where >
Pdx P dx P PB
P dx Pdx P P
(4)
Now, true-to-life coefficients for landscape friction can be calculated by optimizing B, as follows:
set B to 0 (5)
or, at least,
minimize B (6)
The optimization of ( )L can be properly achieved using genetic algorithms (GAs; Holland, 1975). GAs
are powerful evolutionary models with wide potential applications in ecology and biology, such as
optimization of protected areas (Ferrarini et al., 2008; Parolo et al., 2009), optimal sampling (Ferrarini, 2012a;
Ferrarini, 2012b), optimal detection of landscape units (Rossi et al., 2014) and networks control (Ferrarini,
2011a; Ferrarini, 2013b; Ferrarini, 2013c; Ferrarini, 2013d; Ferrarini, 2013e; Ferrarini, 2014c). At time T0,
(7)
Now, sloping flow connectivity acts upon the optimized frictional landscape of eq. (7) as follows:
0( , , ( ))
( , )sL x y L
v x y
(8)
In other words, sloping flow connectivity calculates (pixel-by-pixel) the slope of the optimized frictional
landscape along the (vectorial) direction v that is a function of x and y dimensions:
v ax by
(9)
For instance, with a= 0 and b= 1 eq (8) calculates the frictional slope toward the N direction; with a = 1 and b=
1, eq. (8) calculates the frictional slope toward the N-E direction. Since a and b are real numbers, sloping
connectivity is able to calculate slope not only along the 8 cardinal directions (N, E, S, O, N-E, S-E, N-O, S-O),
but along any compass bearing.
In order to estimate the slope of a landscape cell along a particular direction, sloping flow connectivity acts as
follows:
0 0( , , ( ))s sL L x y L
125
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
0( , , ( ))
( , )s v
h
L x y L D
Dv x y
(10)
where Dv and Dh are the vertical and horizontal distances between the center-point of the focal cell and the
center-point of the adjacent cell in the specified direction.
Sloping flow connectivity takes a reasonably simple approach to estimate the % slope of a cell in a particular
direction: it assumes that the elevation values for all cells are good estimates of the elevation at the center-
points of the cells. If the direction is not a cardinal direction, then the slope is calculate from the focal cell
center-point to an interpolated point between 2 adjacent cell center-points. It can also be calculated in degrees
using the following alternative equation:
0( , , ( )) 180arctan( )*
( , )s v
h
L x y L D
Dv x y
(11)
Which is the ecological meaning of the slope direction (also known as, slope aspect or slope orientation)
calculated over the frictional landscape Ls? Clearly, it represents the least friction direction (LFD) to species
dispersal. To calculate the aspect from the frictional landscape Ls, I have used the equation from Evans (1972):
0
0
( , , ( ))
arctan( , , ( ))
s
s
L x y L
yLFD
L x y L
x
(12)
which is the angle by the x and y derivative of Ls via arctan, measured clockwise in degrees from North. Once
LFD is calculated, it can be then grouped into 9 categories (flat; N: 337.58–22.58, NE: 22.58–67.58, E:
67.58–112.58, SE: 112.58–157.58, S: 157.58–202.58, SW: 202.58–247.58, W: 247.58–292.58, NW:
292.58–337.58).
In order to apply sloping flow connectivity modelling to real landscapes, I wrote the ad hoc software
Connectivity Lab (Ferrarini, 2013f).
3 An Applicative Example
The Ceno valley is a 35,038 ha wide valley situated in the Province of Parma, Northern Italy. It has been
mapped at 1:25,000 scale (Ferrarini, 2005; Ferrarini et al., 2010) using the CORINE Biotopes classification
system. The landscape structure of the Ceno Valley has been widely analysed (Ferrarini and Tomaselli, 2010;
Ferrarini, 2011b; Ferrarini, 2012c; Ferrarini, 2012d).
From an ecological viewpoint, the most interesting event registered in the last years is the shift of wolf
populations from the montane belt to the lowland. Several populations have been recently observed in situ by
life-watchers, environmental associations and local administrations.
As an example of sloping flow connectivity, I have applied my model to a portion of the Ceno valley (Fig.
1) above 1000 m a.s.l. close to the municipality of Bardi where several small populations of wolves have been
recently observed. The area is a square of about 20 km * 20 km. Optimized friction values to wolf
presence are borrowed from Ferrarini (2012e) in the form of friction coefficients assigned to every land cover
classes. A discussion of wolf’s frictional coefficients is outside the goals of this paper, so I avoid presenting
them.
( )L
126
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
Fig. 1 The frictional landscape Ls has been built for wolf upon a 20 km * 20 km portion of the Ceno Valley (province of Parma, Italy) that represents here the real landscape L(x,y,z,t). The higher frictional values are in red, the lower ones are in blue. The frictional landscape has been built using both structural and functional properties of the landscape (Ferrarini, 2012e).
In order to simulate (directional) species movements along such landscape, I calculated sloping flow
connectivity along the 4 cardinal directions (N, S, E, O) using equations from (8) to (12).
Simulated northward (i.e., a= 0 and b= 1) connectivity is depicted in Fig. 2. Darker areas depict the
individuated barriers, lighter ones are the detected facilities. Barriers to northward wolf’s movements are
present in particular in the Eastern and Western portions of the study area. Instead, the central part of the study
area present less problems to northward wolf’s movements.
Simulated eastward (i.e., a= 1 and b= 0) connectivity is depicted in Fig. 3. It is clear that the study area is
particularly knotty with regard to eastward movements. Only the central portion allows for facilitated eastward
shifts. Southward (a= 0, b= -1) and westward (a= -1, b= 0) movements are simulated in Fig. 4 and Fig. 5.
Barriers and facilities are clearly detected.
127
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
Fig. 2 Simulated wolf’s northward movements (as indicated by the arrow). Darker areas represent the individuated barriers, lighter ones are the detected facilities.
Fig. 3 Simulated wolf’s eastward movements (as indicated by the arrow). Darker areas represent the individuated barriers, lighter ones are the detected facilities.
128
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
Fig. 4 Simulated wolf’s southward movements (as indicated by the arrow). Darker areas represent the individuated barriers, lighter ones are the detected facilities.
Fig. 5 Simulated wolf’s westward movements (as indicated by the arrow). Darker areas represent the individuated barriers, lighter ones are the detected facilities.
129
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
The computation of the LFD for the frictional landscape Ls using eq. (12) depicts for each portion of the
study area the most facilitated direction for wolf (Fig. 6).
It results clear that wolf’s movements towards South, South-West and West are the most probable (i.e.
facilitated) events in the study area. Instead, wolf’s movements towards North, North-East and East are highly
improbable. It’s also clear that there’s a kind of spatial clustering of movements probabilities, with certain
directions prevalent in particular portions of the study area (Fig. 6).
Fig. 6 Map of the most facilitated wolf’s movements in the study area. It has been achieved via sloping flow connectivity applied to the frictional landscape Ls of Fig. 1. For each pixel, the direction with the lowest friction to species dispersal is given.
The overall assessment of the least friction directions to wolf’s dispersal in the study area is given in Fig.
7, which refers to Fig. 6.
130
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
Fig. 7 Histogram of the most facilitated directions to wolf’s dispersal in the study area. Columns count the number of pixels in the sloping landscape of Fig. 6 with a particular direction for the most facilitated wolf’s dispersal.
4 Conclusions
Planning greenways and ecological networks in an effective manner is not an easy task, since it requires to
detect the real potential of each landscape elements to facilitate or obstruct both directional and overall species
movements. Conservation practitioners use two main strategies to promote connectivity. The first focalizes on
conserving areas that facilitate movement, the second focuses on restoring connectivity across areas that
impede movement. Most connectivity works have focused on the former strategy.
In this paper, I have introduced sloping flow connectivity that is on top of this requirement, as it is able to
produce simple and effective maps of barriers and facilities to directional and overall species dispersal.
Sloping flow connectivity takes advantage of two previously-introduced theoretical and methodological
frameworks for the prediction of species dispersal: flow connectivity and reverse flow connectivity.
References
Couvet D. 2002. Deleterious effects of restricted gene flow in fragmented populations. Conservation Biology,
16: 369-376
Evans I.S. 1972. General geomorphometry, derivations of altitude and descriptive statistics. In: Chorley R.J.
(Editor). Spatial analysis in geomorphology. 17-90, Harper & Row Publishers, New York, USA
Fahrig L, Merriam G. 1985. Habitat patch connectivity and population survival. Ecology, 66: 1762-1768
Ferrarini A. 2005. Analisi e valutazioni spazio-temporale mediante GIS e Telerilevamento del grado di
Pressione Antropica attuale e potenziale gravante sul mosaico degli habitat di alcune aree italiane. Ipotesi
di pianificazione. Ph.D. Thesis, Università degli Studi di Parma, Parma, Italy
Ferrarini A, Rossi G, Parolo G, Ferloni M. 2008. Planning low-impact tourist paths within a Site of
Community Importance through the optimisation of biological and logistic criteria. Biological
Conservation, 141:1067-1077
Ferrarini A, Bollini A, Sammut E. 2010. Digital Strategies and Solutions for the Remote Rural Areas
Development. Acts of Annual MeCCSA Conference, London School of Economics, London, UK
131
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
Ferrarini A, Tomaselli M. 2010. A new approach to the analysis of adjacencies. Potentials for landscape
insights. Ecological Modelling, 221:1889-1896
Ferrarini A. 2011a. Some thoughts on the controllability of network systems. Network Biology, 1(3-4): 186-
188
Ferrarini A. 2011b. Network graphs unveil landscape structure and change. Network Biology, 1(2): 121-126
Ferrarini A. 2012a. Biodiversity optimal sampling: an algorithmic solution. Proceedings of the International
Academy of Ecology and Environmental Sciences, 2(1): 50-52
Ferrarini A. 2012b. Betterments to biodiversity optimal sampling. Proceedings of the International Academy
of Ecology and Environmental Sciences, 2(4): 246-250
Ferrarini A. 2012c. Founding RGB Ecology: the Ecology of Synthesis. Proceedings of the International
Academy of Ecology and Environmental Sciences, 2(2): 84-89
Ferrarini A. 2012d. Landscape structural modeling. A multivariate cartographic exegesis. In: Ecological
Modeling (Zhang WJ, ed). 325-334, Nova Science Publishers Inc., New York , USA
Ferrarini A. 2012e. The ecological network of the province of Parma. Province of Parma, Parma, Italy (in
Italian)
Ferrarini A. 2013a. A criticism of connectivity in ecology and an alternative modelling approach: Flow
connectivity. Environmental Skeptics and Critics, 2(4): 118-125
Ferrarini A. 2013b. Controlling ecological and biological networks via evolutionary modelling. Network
Biology, 3(3): 97-105
Ferrarini A. 2013c. Computing the uncertainty associated with the control of ecological and biological systems.
Computational Ecology and Software, 3(3): 74-80
Ferrarini A. 2013d. Exogenous control of biological and ecological systems through evolutionary modelling.
Proceedings of the International Academy of Ecology and Environmental Sciences, 3(3): 257-265
Ferrarini A. 2013e. Networks control: Introducing the degree of success and feasibility. Network Biology, 3(4):
127-132
Ferrarini A. 2013f. Connectivity-Lab 2.0: a software for applying connectivity-flow modelling. Manual, Italy
(in Italian)
Ferrarini A. 2014a. True-to-life friction values in connectivity ecology: Introducing reverse flow connectivity.
Environmental Skeptics and Critics, 3(1): 17-23
Ferrarini A. 2014b. Can we trace biotic dispersals back in time? Introducing backward flow connectivity.
Environmental Skeptics and Critics, 3(2): 39-46
Ferrarini A. 2014c. Local and global control of ecological and biological networks. Network Biology, 4(1): 21-
30
Gardner RH, Gustafson E.J. 2004. Simulating dispersal of reintroduced species within heterogeneous
landscapes. Ecological Modelling, 171: 339-358
Gustafson EJ, Gardner R.H. 1996. The effect of landscape heterogeneity on the probability of patch
colonization. Ecology 77, 94–107
Hanski I, Ovaskainen O. 2000. The metapopulation capacity of a fragmented landscape. Nature, 404: 755-758
Holland J.H. 1975. Adaptation in Natural And Artificial Systems: An Introductory Analysis with Applications
to Biology, Control and Artificial Intelligence. University of Michigan Press, Ann Arbor, USA
Kareiva P, Wennergren U. 1995. Connecting landscape patterns to ecosystem and population processes.
Nature, 373: 299-302
Kennedy M, Gray R.D. 1997. Habitat choice, habitat matching and the effect of travel distance. Behaviour,
134: 905-920
132
Proceedings of the International Academy of Ecology and Environmental Sciences, 2014, 4(3): 123-133
IAEES www.iaees.org
King AW, With KA. 2002. Dispersal success on spatially structured landscapes: when do dispersal pattern and
dispersal behaviour really matter? Ecological Modelling 147: 23-39
McRae BH. 2006. Isolation by resistance. Evolution, 60: 1551-1561
McRae BH, Beier P. 2007. Circuit theory predicts gene flow in plant and animal populations. Proceedings of
the National Academy of Sciences of the USA, 104: 19885-19890
McRae BH, Dickson BG, Keitt TH, Shah VB. 2008. Using circuit theory to model connectivity in ecology and
conservation. Ecology, 10: 2712-2724
Parolo G, Ferrarini A, Rossi G. 2009. Optimization of tourism impacts within protected areas by means of
genetic algorithms. Ecological Modelling, 220: 1138-1147
Rossi G, Ferrarini A, Dowgiallo G, Carton A. et al. 2014. Detecting complex relations among vegetation,
soil and geomorphology. An in-depth method applied to a case study in the Apennines (Italy).
Ecological Complexity, 17(1): 87-98
Tischendorf L. 1997. Modelling individual movements in heterogeneous landscapes: potentials of a new
approach. Ecological Modelling 103: 33-42
Tischendorf L, Fahrig L. 2000. How should we measure landscape connectivity? Landscape Ecology 15, 633–
641
Wiegand T, Moloney K, Naves J, Knauer F. 1999. Finding the missing link between landscape structure and
population dynamics: a spatially explicit perspective. American Naturalist, 154: 605-627
With KA, Gardner RH, Turner M.G. 1997. Landscape connectivity and population distributions in
heterogeneous environments. Oikos, 78: 151-169
133
Proceedings of the International Academy of Ecology and
Environmental Sciences
The Proceedings of the International Academy of Ecology and Environmental Sciences (ISSN 2220-8860) is an open access, peer-reviewed online journal that considers scientific articles in all different areas of ecology and environmental sciences. It is the flagship journal of the International Academy of Ecology and Environmental Sciences. It dedicates to the latest advances in ecology and environmental sciences. The goal of this journal is to keep a record of the state-of-the-art research and promote the research work in these fast moving areas. All manuscripts submitted to Proceedings of the International Academy of Ecology and Environmental Sciences must be previously unpublished and may not be considered for publication elsewhere at any time during review period of this journal. The topics to be covered by this journal include, but are not limited to theory, methodology, technology, innovation, activity, and project in the following areas:
Animal ecology, plant/microbe ecology, wetland ecology, farmland ecology, forest/grassland ecology, marine ecology, pollution ecology, etc.
Biological conservation & preservation, ecosystem restoration, environmental policy, environmental toxicology, environmental pollution and control, natural resource, bioenergy research, environmental technology, waste management, environmental economics, environmental management & planning, environmental education, environmental engineering, global climate change, oceanography, etc.
Authors can submit their works to the email box of this journal, [email protected].
In addition to free submissions from authors around the world, special issues are also accepted.
The organizer of a special issue can collect submissions (yielded from a research project, a
research group, etc.) on a specific topic, or submissions of a conference for publication of special
issue.
Editorial Office: [email protected]
Publisher: International Academy of Ecology and Environmental Sciences
Address: Flat C, 23/F, Lucky Plaza, 315-321 Lockhart Road, Wanchai, Hong Kong
Tel: 00852-6555 7188
Fax: 00852-3177 9906
E-mail: [email protected]
Proceedings of the International Academy of Ecology and Environmental Sciences ISSN 2220-8860 Volume 4, Number 3, 1 September 2014 Articles
Morphometry and meristic counts of Bombay duck, Harpodon nehereus
(Hamilton, 1822) along Sunderban region of West Bengal, India
V. V. Kumar, A. D. Reddy, S. R. Choudhury, et al. 95-105
Classic models of population dynamics: assumptions about selfregulative
mechanisms and numbers of interactions between individuals
L.V. Nedorezov 106-113
Multivariate statistical analysis of surface water chemistry: A case study of
Gharasoo River, Iran
MH Sayadi , A Rezaei1, MR Rezaei, K Nourozi 114-122
Detecting barriers and facilities to species dispersal: Introducing sloping flow
connectivity
Alessandro Ferrarini 123-133
IAEES http://www.iaees.org/