Xing Fang1, Ding Longjiang2, Zhao Liang2 and Wan-Ran Zhang3
1Associate Professor at the Department of Civil Engineering, Lamar University,
Beaumont, Texas 77710-0024, USA, (409) 880-2287, (409) 880-8121 (fax), fangxu@hal.lamar.edu
2Research Assistants at the Department of Civil Engineering, Lamar University,
Beaumont,
Texas 77710-0024, USA, (409) 880-8447(409), 880-8121 (fax)
3Professor at the Department of Computer Science, Lamar University, Beaumont,
Texas 77710-0056, USA, (409) 880-8748, (409) 880-2364 (fax), zhang@hal.lamar.edu
Abstract: A modeling
system (URL http://lakefish.lamar.edu/), which can be accessed by any user via
the Internet (World Wide Web), is developed to make water quality and
fish habitat projections in small lakes (up to 20 km2). The
projections require complex numerical models that integrate various physical and
biochemical processes such as hydrodynamics, air-water surface exchange, and
biochemical reactions. Basic concepts and methods for model projections are
summarized. The system is designed for three
user levels: the general public, inexperienced researchers, and experienced
water quality modelers. The system is based on the three-tier web
application architecture, which is divided into the client tier (web browser),
the server tier (web server with servlet engine), and the database tier. The database tier currently stores information
about weather conditions from 1961 to 1995 over 243 locations in the continental
United States and a future climate scenario due to doubling of atmospheric CO2
concentration. This study is to explore means by which advanced computing
technology can facilitate software reuse, data sharing, and decision support
through the Internet in the context of ecosystem management.
Keywords: web-based computing, lake water quality models, client/server technology
Protecting the
environmental quality of aquatic ecosystems in the U.S.A. and the world is of
great interest and importance to the public and the global economy. Water
quality protection and restoration require field data collection and water
quality projection using numerical models. The projections provide vital
information to ecosystem managers and researchers (Jorgensen et al., 1995),
especially for regional management practice and risk assessment under future
climate conditions (e.g., possible doubling of CO2 concentration in
the atmosphere) (Fang, 1997; Stefan et al. 1996). About 1000 models for
environmental and ecological applications have been described and referenced in
the literature during the last decade (Jorgensen et al., 1995). Since most of
the models were designed for a specific ecosystem and tailored to a specific
problem, extensive modeling experience and model modification are required in
order to solve new environmental problems using the existing models.
One example of the above models is a year-round Lake Water Quality and Fish Habitat projection (LWQFH) system (Fang, 1997; Stefan et al. 1996). The system was developed, calibrated, and verified against measured data (over 5000 data sets) for examining the impact of climate variability on water quality and fish habitat in small lakes of the contiguous U.S. (Stefan et al., 1996; Fang et al., 1999). The applications and development of the system were continuously supported by the U.S. Environmental Protection Agency (U.S. EPA). The system has been successfully applied to lakes/reservoirs in Texas, U.S.A. (Fang, 1997). Currently this modeling system is a FORTRAN legacy program with over 4,000 lines of source code, since it must integrate various physical and biochemical processes in lakes, e.g. hydrodynamics (inflow and wind mixing), air-water surface exchange, and biochemical reactions (Fig. 1). The system is complex and currently can only be used in specific research groups by experienced users.
In order to allow many water-quality and fishery scientists/managers in Texas and the U.S. to utilize the LWQFH system, there are several reasonable approaches to convert the system into an Internet tool. For example, (1) Internet-based distribution, which provides executable and/or source code of the system, and a users manual on the Internet for free download (e.g., many software packages available from USEPA); (2) Internet-shared computing environment, where instructions and a data submission form are provided on the Internet and users initiate a program run, but the program is executed on the provider’s machine (server) and a report of simulation results is returned to the client. This paper summarizes a web-based (the approach 2) modeling system for water quality and fish habitat projections in lake. The system has been designed for three user levels: the general audience, inexperienced water-quality modelers, and experienced ones. It will enable inexperienced users (e.g. high school students in science classes) to perform tasks that were considered otherwise impossible and will assist experienced users to accomplish more difficult modeling tasks more effectively and efficiently.
Fig. 1 A schematic representation of some physical processes occurring within a lake (Riley and Stefan, 1987).
A
process-oriented, unsteady, one-dimensional (vertical) year-round lake water
quality model MINLAKE96 (Fang and Stefan, 1996, 1997; Stefan et al., 1998),
which simulates daily water temperature and DO profiles as well as ice/snow
covers on lakes in a continuous mode for multiple years, was used in this study.
The year-round simulated daily water temperature and DO simulations were tested
against water temperature and DO measurements at different water depths (5,976
data pairs) for a total of 48 ‘lake-years’ (Fang and Stefan, 1996, 1997;
Stefan et al., 1998). Average standard error between simulated and measured
water temperatures year-round was 1.4 oC, for the ice-cover season
alone it was 0.5 oC (Fang and Stefan, 1996). Average standard error
between year-round simulations and measurements of DO concentrations was 1.94
mg/liter. Simulated ice thicknesses and snow depths were tested against
measurements in three lakes (128 data pairs over 8 years). Standard errors
between simulated and measured values were 0.07 m for snow depths and 0.12 m for
ice thicknesses (Fang and Stefan, 1996). Predicted freeze-over dates were
compared with observations in nine Minnesota lakes for multiple (1 to 36) years
(Fang et. al, 1996). The difference between the simulated and observed ice
formation dates was less than 7 days for all lakes studied.
The year-round
numerical simulation model for daily water temperature profiles in a lake solves
the one-dimensional, unsteady heat transfer equation
(1)
where Tw(z, t) is the daily water temperature (oC) as a function of depth z (m) and time t (day), A(z) is horizontal lake area (m2) also a function of depth z (m), KZ(z, t) is the vertical turbulent heat diffusion coefficient (m2 day-1), r is the density of water, cp is heat capacity of water, and Hw is the net internal heat source or sink strength per unit volume (J m-3 day-1). Solar radiation absorption in the water column is the main contributor to the heat source term. Heat exchange between the lake and the atmosphere is treated as a source/sink term for the topmost water layer of a lake during the open water season. It includes short-wave solar radiation, long-wave atmospheric solar radiation, evaporation and convection through the water surface, and back radiation from water. The MINLAKE96 model includes also heat exchange between each water layer and its adjoining sediments (Fang et. al, 1996). Submodels for snow and ice thickness simulations were originally developed by Gu and Stefan (1990) but have been used with some modifications (Fang and Stefan, 1996). A new physical-based algorithm, which replaces the previous empirical and lake size dependent criteria for the date of ice formation, was developed and incorporated in the model (Fang et. al, 1996). During the ice cover period in cold regions, the model simulates ice thicknesses and sediment temperature profiles (heat conduction equation) first, then determines the heat source/sink term, and finally solves the heat transfer equation (1) to obtain water temperature profiles below the ice. The model uses a stacked layer system, the layers consisting of lake sediments, water, ice cover and snow cover (Gu and Stefan 1990; Fang and Stefan 1996). Snow thickness is determined from snow accumulation, followed by compaction and melting of snow by surface heat input (convection, rainfall, solar radiation) and melting within the snow layer due to internal absorption of short wave radiation, and transformation of wetted snow to ice when cracks in the ice cover allow water to spill onto the ice surface (Fang and Stefan, 1996). In the model ice growth occurs from the ice/water interface downward and from the ice surface upward (Fang and Stefan, 1996). Ice decay occurs at the snow/ice interface, ice/water interface, and within the ice layer.
The year-round
numerical simulation model for daily dissolved oxygen profiles in a lake solves
the one-dimensional, unsteady transport equation (Stefan et al., 1993)
(2)
where C(z, t) is the DO concentration in mg/liter as a function of depth (z)
and time (t), Sb is the coefficient for sedimentary oxygen demand (SOD)
in mg O2/m2/day,
PMAX is the maximum specific
oxygen production rate by photosynthesis at saturating light conditions in mg
O2/(mg Chla)/day, Min[L] is the light limitation determined by Haldane
kinetics, Chla is the chlorophyll-a
concentration in mg/liter, YCHO2 is the yield coefficient, i.e. the ratio of mg chlorophyll-a
to mg oxygen, kb and kr
are the first order decay rate coefficients for BOD and plant respiration (per day), respectively, qb and qr are the temperature adjustment coefficients for BOD and plant respiration,
BOD is the biochemical oxygen demand
concentration in mg/liter. In the model, the oxygen transfer through the water
surface (reaeration) during the open water season is expressed as
, and is used as an oxygen source or sink term in the topmost water (surface)
layer of the lake. Cs,
ke and
are the saturated DO concentration at the surface water temperature, surface
oxygen transfer coefficient calculated as a function of windspeed, and thickness
of the top most water layer, respectively. Diffusive oxygen flux at the lake
bottom is set equal to zero as a boundary condition. Sedimentary oxygen demand
is treated as a source/sink term for each layer, since each layer is in contact
with sediments.
For the DO
simulations in a lake during the winter ice cover period, modifications had to
be made in Equation (2) to account for the presence of an ice cover and low
temperatures. These modifications include: (a) reaeration is zero (ke
is set equal to zero) because the lake ice cover prevents any significant gas
exchange between the atmosphere and the water body; (b) oxygen consumption by
plant respiration is very small and is not presented as a separate sink term (kr = 0.0) because in winter,
chlorophyll-concentrations are fairly small and water temperatures are very low;
(c) water column oxygen demand, WOD, by detritus and other matter, is set
constant (0.01 g O2/m3/day) regardless of trophic status
of a lake (Mathias and Barica, 1980); and (d) sedimentary oxygen demand (Sb) is made dependent on
trophic state and set equal to 0.226, 0.152, and 0.075 (g O2/m2/day)
for eutrophic, mesotrophic, and oligotrophic lakes, respectively, based on field
studies summarized by Barica and Mathias (1979) in ice covered temperate zone
lakes in Canada. DO concentrations were simulated after water temperature and
snow/ice covers had been simulated. Equations (1) and (2) were solved
numerically for time steps of one day and layer thicknesses of one meter using
an implicit finite difference scheme and a Gaussian elimination method.
Fish habitat for
three fish guilds was estimated from simulated daily water temperature and
dissolved oxygen profiles in lakes. Thermal fish habitat could be estimated from
measured water temperature and DO profiles (e.g., Christie and Regier 1988).
Fish thermal and DO criteria for survival and good-growth (Table 1) were applied
to the simulated daily water temperature and DO profiles as shown schematically
in Figure 1. This figure is for a lake where
fish cannot be present during much of the open-water season because low DO
concentrations near the lake bottom in summer force fish upward, and warm water
temperatures near the water surface force fish downward in search of conditions
for survival. When isotherm of lethal temperature (LT) and the DO limit isopleth
for a fish species intersect each other, uninhabitable space exists over the
entire depth of a stratified lake. Summerkill of a fish species in a lake is
expected to occur if these conditions last more than seven days. When the
maximum surface water temperature is lower than the LT, the isotherm for LT will
not show up. In this case the DO survival limit becomes the only survival
criterion, especially during the winter ice cover period. The DO survival limit
(isopleth) can occur not only during the open-water season (summer) but also in
the ice-cover period (winter). Other isotherms designate the survival
temperature (LT), the upper good-growth temperature (UGGT) and the lower
good-growth temperature (LGGT). Between these isotherms, three fish habitats are
identified:
(1) Uninhabitable space if temperature is above or DO is below the survival limit.
(2) Good growth habitat if temperature is between the upper and lower good-growth limits (i.e. LGGT < T < UGGT) and DO is above the survival limit.
(3) Restricted growth habitat if temperature is above the upper good growth temperature but below the survival limit (i.e. UGGT < T <LT), or if temperature is below the lower good- growth limit of temperature (i.e. T < LGGT) and DO is above the survival limit.
To quantify whether
fish habitat can exist in a lake, three parameters (NSB, NSE, and NSL) shown in
Figure 1 were extracted from the depth-time contours of fish habitat. NSB and
NSE indicate the beginning and the end dates of non-survival at all depths in a
lake, respectively. NSB and NSE can occur during the open-water season or during
the ice cover season. These NSB and NSE values bracket the fish “summerkill”
or “winterkill” period of length NSL, the latter primarily due to low DO
values. NSL is the total number of days when either temperature or dissolved
oxygen does not meet the fish presence criteria at all depths of a lake. NSL
defines the length of summer or winter periods of non-survival or the length of
time that fish presence criteria are not met. NSL was determined as the
difference between NSE and NSB for the winter and the summer seasons,
separately. In this study, when NSL is greater than 7 days during the open water
season, “summerkill” is assumed to occur, and when NSL is greater than 7
days during the winter ice cover period, “winterkill” is assumed to occur.
Low water temperature, e.g. at or below 2oC, may also cause fish
mortality through osmoregulatory dysfunction (Johnson and Evans 1996).
"Winterkill" due to low temperatures is not examined in this study,
and low DO is the only factor to control winterkill in lakes.
Fig. 2 Schematic distribution over time and depth of those isotherms and dissolved oxygen isopleths relevant to the survival and growth of a fish species, in a seasonally stratified lake. Fish survival and good-growth parameters defined are shown (after Hondzo and Stefan 1996). LT = lethal temperature isotherm, UGGT = upper good-growth temperature, LGGT = lower good-growth temperature, NSB = non-survival beginning date, NSE = = non-survival ending date, NSL = = non-survival length (period).
A web site (http://lakefish.lamar.edu/) has been developed to provide access to the existing LWQFH system via the Internet. The system provides forms for submitting input data in order to run the models and reporting simulation results via the Internet. Input data include simulation control parameters, lake geometry, model coefficients, initial and boundary conditions. Several databases were developed in order to identify the nearest weather stations for running the lake models. Databases contain daily weather records at 243 stations in the contiguous United States. By accessing the web site via the Internet, users can get a better understanding of the modeling system and of the status and variability of water quality and fish habitat in freshwater ecosystems under present and projected climates (e.g., a doubling of atmospheric CO2).
The web site is running under a Windows NT Server and the Microsoft IIS (Internet Information Server) web server. The web application for the modeling system of lake water quality and fish habitat is based on the three-tier web application structure (Fig. 3), which is divided into three layers: client (web browser), server (web server plus servlet engine), and database. This three-tier structure (model) separates information presentation, business logic, and database, and makes the web application more robust and easy to maintain. In Fig. 3, HTTP stands for Hyper-Text Transfer Protocol as a primary network protocol for the Internet, IIS stands for the Internet Information Server that is the Microsoft web-server on Window NT and Window 2000. JVM (Fig. 3) stands for Java Virtual Machine as an independent virtual operating system built on the native OS (Operating System) that makes the Java language platform independent. Information request and/or exchange between client and server (e.g. database) processes through Javax.servlet (Servlet Engine). All databases for this modeling system were developed and managed under Microsoft Access. JDBC stands for Java DataBase Connectivity, a uniform java interface, which stands between Java language and underlay variable databases. With JDBC, it is possible “written once, use everywhere”, which means that regardless of different databases (e.g., Oracle, Sybase, MS Access, MS SQL) the application programmers developed can use the uniform embedded-SQL pattern independent of underneath database platforms. On the client site, any Internet user can use a web browser (e.g. MS Internet Explorer, Netscape Communicator) to access information for the modeling system, which consists of a series of regular HTML or XML (dynamic) web pages with enhanced features via Java Applets, JavaScript or VBScript (Visual Basic). The FORTRAN legacy program LWQFH runs through the Java Servlet Engine.
Fig. 3 Three-tier web application structure with Java Servlet for the modeling system of lake water quality and fish habitat.
Fig. 4 A graphic representation of simulated and measured water temperatures in the modeling system.
A modeling system (URL http://lakefish.lamar.edu/), which can be accessed by any user via the Internet (World Wide Web), was developed to make water quality and fish habitat projections in small lakes (up to 20 km2). The projections require complex numerical models that integrate various physical and biochemical processes such as hydrodynamics, air-water surface exchange, and biochemical reactions. Basic concepts and methods for water quality and fish habitat projections have been summarized. The system provides forms for submitting input data in order to run the models and reporting simulation results via the Internet. The system was designed for three user levels: the general public, inexperienced researchers, and experienced water quality modelers. The system is based on the three-tier web application structure, which is divided into three layers: client (web browser), server (web server with servlet engine), and database. The system includes several databases which store information about weather conditions from 1961 to 1995 over 243 locations in the continental United States and a future climate scenario due to doubling of atmospheric CO2 concentration. This study shows that advanced computing technology can facilitate software reuse, data sharing, and decision support through the Internet in the context of ecosystem management.
Acknowledgements
This material is based in part upon work supported by the Texas Advanced Research (Technology) Program under Grant No. 003581-0005-1999 and the U.S. Environmental Protection Agency, Office of Research and Development.
Christia, C.G., Regier, H.A., 1988. Measurements of optimal habitat and their relationship to yields for four commercial fish species. Can. J. Fish. Aquat. Sci., 45, 301-314.
Fang, X., Stefan, H.G., 1996. Development and validation of the water quality model MINLAKE96 with winter data. Project Report 390, St. Anthony Falls Laboratory, University of Minnesota, Minneapolis, MN.
Fang, X., Ellis, C.E., Stefan, H.G., 1996. Simulation and observation of ice formation (freeze-over) in a lake. Cold Regions Science and Technology, 24, 129-145.
Fang, X., 1997. “Water Temperature Dynamics in Texas Lakes”. Proceedings of Texas Section of American Society of Civil Engineers - 1997 Spring Meeting, Houston, Texas.
Fang, X., Stefan, H.G., 1997. Simulated climate changes on dissolved oxygen characteristics in ice-covered lakes. Ecological Modelling, 103(1997), 209-229.
Fang, X., H.G. Stefan, and S.R. Alam, 1999. Simulation and validation of fish thermal habitat in north-central U.S. lakes under different climate scenarios. Ecological Modelling, 118:167-191.
Gu, R., Stefan, H.G., 1990. Year-round temperature simulation of cold climate lakes. Cold Regions Science and Technology 18(2), 147-160.
Johnson, T.B., Evans, D.O., 1996. Temperature constraints on overwinter survival of age-o White Perch. Trans. Am. Fish. Soc., 125, 466-471.
Jorgensen, S.E., Halling-Sorensen, B., and Nielsen, S.N. (editors), 1995. “Handbook of environmental and ecological modeling”, CRC Lewis Publishers, 672pp.
Mathias, J.A., Barica, J., 1980. Factors controlling oxygen depletion in ice-covered lakes. Canadian Journal of the Fisheries Aquatic Sciences 37:185-194.
Stefan, H.G., Hondzo, M., Fang, X., 1993. Lake water quality modeling for projected future climate scenarios. Journal of Environmental Quality, 22(3), 417‑431.
Stefan, H. G., M. Hondzo, X. Fang, J. G. Eaton, and J. H. McCormick, 1996. “Simulated longterm temperature and dissolved oxygen characteristics of lakes in the north-central United States and associated fish habitat limits.” Limnology and Oceanography, 41(5):1124-1135.
Stefan, H.G., Fang, X., Hondzo, M., 1998.
Simulated climate change effects on year-round water temperatures in temperate
zone lakes. Climatic Changes, 40:547-576.