3Reseach Assistant at the Department of Civil Engineering, Lamar University,

Beaumont, Texas 77710-0024, USA, (409) 880-8447

Abstract: Sabine Lake is a relatively shallow, microtidal estuary located in the southeast of Texas, USA, and directly connected to the Gulf of Mexico. This study is to examine interactions between saltwater and freshwater inflows on Sabine Lake system. Salinity, water temperature, water level, and dissolved oxygen measurements were collected continuously every hour in both Johnson and Black Bayous of Sabine Lake. Water samples were collected monthly and analyzed in the laboratory to measure five water quality parameters (Total Kjeldahl nitrogen, nitrite plus nitrate, ammonium nitrogen, total phosphorus, and total organic carbon). The data were analyzed to provide information on movement of saltwater and nutrients moving to and from Sabine Lake and the marsh lands on the lake’s east side.

 

Keywords: estuary, water quality, nutrient budget, and monitoring.

The region between the free flowing river and the ocean is a fascinating, diverse, and complex water system: the coastal regime of estuaries, bays and harbors. The incursion of salinity from the ocean and the influx of nutrients from the upstream drainage contribute to the generation of a unique aquatic system. The exchange of nutrients from marshland (wetland) can be an important component for nutrient budget in estuaries. The Sabine-Neches estuary studied is located in the southeast Texas, USA, and directly connected to the Gulf of Mexico. It receives discharge from the Neches and Sabine rivers, as well as from local watersheds. The estuary contains the second-largest freshwater marsh along the Texas coast, the largest salt marsh, and the largest bottomland hardwood swamps. The marsh production provides a major source of organic material supporting the estuary. The marsh may assimilate substantial volumes of municipal and industrial waste and convert them into organic material.

One model of estuarine production has held that production from deltaic marshes feeds productivity throughout the estuary. Anecdotal evidence from TPWD (Texas Parks and Wildlife Department) Coastal Fisheries monitoring teams suggests that the side of Sabine Lake bordering Louisiana marshes is biologically richer than the Texas side. Therefore it may be that marsh production is more important to Sabine Lake productivity than is material or nutrients brought in by Sabine and Neches Rivers. This study provides data and analysis to characterize the magnitude of marsh contribution. This bayou study will provide data for estimating nutrient contribution from east-side marshes to Sabine Lake. A series of seasonal water quality and flow data were collected from bayous on the Louisiana side of the Lake where there is now little or no data. Data from this study will feed into a future analytical modeling project and are required for a complete nitrogen budget of Sabine Lake. The Sabine Lake nitrogen budget will allow an assessment of how much Sabine Lake depends on nutrient contributions from the Sabine and Neches rivers and from other sources, such as discharges from municipal wastewater treatment plants.

The specific objectives of this study are to (1) monitor salinity and other water quality parameters in two bayous connecting Sabine Lake to marshland on the east side; and (2) to collect data on nutrient transport in two bayous (Johnson Bayou and Black Bayou). Salinity monitoring was accomplished through maintenance of instruments to continuously measure and record water quality. Salinity monitoring enables relationships to be established between inflows and bayou salinities, for model calibration. Nutrient transport was derived from synoptic measurements of flows and dissolved and suspended nutrients in both bayous.

The monitoring task includes monthly site visits to exchange instruments, routine calibration and instrument maintenance in Black Bayou and in Johnson Bayou of Sabine Lake. Water level, salinity and two water quality parameters (water temperature, dissolved oxygen concentration) were monitored continuously with automated recorders (Hydrolab Datasondes). The instruments were programmed to record water quality measurements every hour, and were individually calibrated with conductivity standards that have salinity values near the range of values expected at each site. The instruments were retrieved from their field installations about every 30 days and replaced with freshly serviced and calibrated units in order to continuously take readings in the coming month. Raw data of the study are available via Internet (http://ceserver.lamar.edu/ fang/rawdata.html).

In an estuary, bio-fouling of the water-quality probes can render measurements invalid within a few weeks; therefore a decalibration procedure was used for provision of reliable data. Decalibration consists of recording the Hydrolab Datasonde measurements of standard solutions as soon as possible after retrieval, before probes or batteries are serviced. This establishes the instrument’s “drift” and validity of the data, or at least provides some information on the degree to which bio-fouling has affected the probe readings. If decalibration could not be done, the monitoring data’s validity is established by comparison with field check data collected at the time of the instrument’s retrieval.

During each monthly field trip, upon arriving at each field site, water quality data were collected at the approximate depth of the monitoring instrument’s deployment -- temperature, conductivity, salinity, and dissolved oxygen--with a calibrated YSI 6920 multiparameter water quality probe, in order to obtain an independent check on the Datasonde record.

The clock time of collecting field-check YSI data was occasionally some minutes different from the recording times of the Datasonde. However, this did not affect comparibility. In comparison of the data measured by YSI and by Datasonde at the closet last-hour, the data show excellent agreement in water temperature (^oC), good agreement in salinity (ppt), and not so good agreement in dissolved oxygen concentration (mg/L) measurements. When we compared DO measured by newly calibrated Hydrolab Datasonde with YSI reading, they were typically in agreement. Since Datasonde stayed in field over one month, especially most of time having low flow conditions, it was found that dissolved oxygen concentrations measured by Datasondes at the end of each one-month period were significantly lower than DO values measured by YSI and calibrated Datasondes just after installation. This may indicate that dissolved oxygen concentrations measured by Hydrolab Datasonde are only reliable within first one to two weeks after installation. Figure 1 shows hourly measurements by YSI and Hydrolab installed side by side at Black Bayou monitoring station in May and June of 2000. Different water levels (Fig. 1) were recorded by the YSI probe and the Hydrolab Datasonde since they used the different datum internally.

Agreement of measured DO by YSI and Hydrolab (Fig. 1) was good (average on 1 mg/l difference) within the first two weeks after installation. After two weeks, difference of measured DO from YSI and Hydrolab was on the average of 3 mg/l with the maximum difference of 6 mg/l (Fig. 1). Technically YSI probe may give more accurate reading of DO under low flow conditions. Hydrolab Datasonde measures DO with a steady state method, whereas YSI uses a rapid pulse method to measure DO intermittently (YSI, 1998). DO sensor is automatically turned on for measuring DO, and then turned off for a period which provides sufficient time to allow DO diffusion towards the DO sensor.

Fig. 1   Measurements from YSI and Hydrolab at Black Bayou of Sabine Lake.

In order to measure dissolved and particulate nutrient flux on incoming and outgoing tides in Johnson and Black Bayous, water quality samples were collected monthly. Analyses were performed according to EPA-accepted QA/QC criteria to measure the following parameters in unfiltered water: (1) Total Kjeldahl nitrogen, TKN; (2) Nitrite plus nitrate, NO₂/NO₃; (3) Ammonium, NH₄, (4) Total phosphorus, TP, (5) Total organic carbon, TOC. Water sampling was collected at both sites. Several field trips were made within 24 hours after Sabine Lake area rainfall events since runoff should wash nutrients from marshlands into bayous and Sabine Lake. Figure 2 shows monthly variation of concentrations of TKN, NO₂/NO₃, and TP at Johnson and Black Bayous. When nutrient concentrations were lower than the detection limits (DL) of analytical equipments, concentrations are shown as empty slots in Figure 2. TKN varies from 0.3 to 0.8 mg/l, and is generally less dependent on rainfall events. High concentrations of TP and NO₂/NO₃ are strongly associated with heavy rainfall events. Ammonia concentrations were typically lower than the detection limit (0.001 mg/l) at both bayous during the study period except May and June of 2000 (Fang, 2000). Average concentrations associated with rainfall events are given in Table 1 for Johnson and Black Bayou when concentrations lower than their detection limits were set to be zero (Gleit, 1985) and were not considered in computing averages. It is assumed that those low concentrations were not related to rainfall events.

Fig.2 Concentrations (mg/l) of TKN, NO₂/NO₃ and TP over time at Johnson and Black Bayous.

Flows and concentrations were combined to produce estimates of materials flux. Total flow to the estuary from drainage basin runoff is found by summing flows originating in both gaged and ungaged watersheds. Gaged flows are obtained from USGS (the United States Geological Survey) streamflow records. Ungaged runoff is computed runoff, using a rainfall-runoff simulation model, based on precipitation over the watershed. Sabine Lake is the watershed 24120 (Fig. 3) in TWDB’s rainfall-runoff simulation model.

Fig. 3 Sabine Lake watersheds for TWDB’s rainfall-runoff simulation model.

The Black Bayou watershed was defined as 0.25 of the 05010 watershed area. The rest of 05010 were summed with Sabine River. The watershed 05008 was split 50:50 between Three bayous and Johnson bayou.

For the purposes of this study, monthly and annual runoff volumes associated rainfall events from September 1999 to August 2000 were estimated from historical flows from 1987 to 1996 as provided by TWDB (Longley, 1994, with recent updates). In order to estimate runoff volume from Johnson and Black Bayous, rainfall data from September of 1999 to August of 2000 were obtained from the Southern Regional Climate Center at the Louisiana State University. Table 2 gives monthly and annual cumulative rainfall in inches from 19987 to 2000 water years. Accumulative rainfall for the 1999 to 2000 water year was very low (38.19 in) in comparison of other years. Efforts were made to match the monthly rainfall in 1999 - 2000 water-year to the monthly rainfall in previous years (1987 to 1996). For example, from Table 2, one can identify that the following months: January of 1988, March of 1989, November of 1989, December of 1996, and December of 1996, had the monthly rainfall very close to the monthly rainfall for September of 1999 (3.25 in). Their daily rainfalls and associated daily flows (runoff) in Johnson Bayou were therefore summarized and compared rainfall event by event (Fang, 2000). The flow in Johnson Bayou during September of 1999 was not simply estimated as the average of the flows from all those months with the same or similar rainfalls. Technical justification (basic rainfall-runoff response concepts) was made to decide which month of rainfall and runoff data would be used. For example, runoff for January of 1988 was too large (7435 cfs) since rainfalls in November of December of 1987 were very large and saturated soils in the watershed, while rainfalls before September of 1999 were not great. For November of 1989, 0.9 inches rainfall did not result any increase of surface flow or runoff, which seems impossible. Finally only flows for March of 1989, December 1989 and December of 1996 were used to take average as estimated flow for September of 1999. The similar process was repeated for other months and for Black Bayou too (Fang, 2000). Results are summarized in Table 3. Estimated total (annual) daily surface runoffs for Johnson Bayou and Black Bayou are 21,729 cfs and 145,208 cfs, respectively. Annual inflows in m³ from marshland into Sabine Lake were derived from the total daily flow (cfs), and are 53x10⁶ and 355x10⁶ m³, for Johnson and Black Bayou, respectively.

Annual nutrient input was estimated from average nutrient concentration and estimated flow from bayou to Sabine Lake due to rainfalls. Results are given in Table 4. Loading is given as 10³ kg or tons per year. In this study, surface runoff was estimated from simulated daily runoff from 1987 to 1996 and rainfall data. Nutrients in Johnson and Black bayous of Sabine Lake are from non-point sources originating in their watersheds. Nutrient inputs are closely related to rainfall events which runoff washes nutrient form marshlands into bayous. In order to more precisely estimate nutrient loading from rainfall storms, one needs a record of flow rate and nutrient concentration during a typical storm in the studying region, and then determines the event mean concentration (EMC) (Wanielista and Yousef, 1992). This was not done in the current work, and needs to be explored in the future study.

Interactions between saltwater and freshwater inflows on Sabine Lake system were examined. Sabine Lake is a relatively shallow, microtidal estuary located in the southeast of Texas, USA, and directly connected to the Gulf of Mexico. Continuous measurements of salinity, water temperature, water level, and dissolved oxygen concentrations were made every hour in two bayous of Sabine Lake system. Water samples were collected monthly and analyzed in the laboratory to measure five water quality parameters (Total Kjeldahl nitrogen, nitrite plus nitrate, ammonium nitrogen, total phosphorus, and total organic carbon). Nutrient concentrations were low during dry periods, but were relatively higher after rainfall events. This indicated that nutrients were washed out by surface runoff from marshland into Sabine Lake. Average concentrations of nutrients associated rainfall and annual surface runoff have been estimated, and have been used to estimate nutrient loading into Sabine Lake for 1999 to 2000 water year.

This study was supported by a grant from the Texas Water Development Board (contract 2000-483-322). Richard McClelland, Ekapoj Trakarrvanvich, and Shoudong Jiang of Department of Civil Engineering at Lamar University participated some of the field trips to collect water samples and to exchange Hydrolab Datasondes. Their supports are greatly appreciated.

Fang, X. 2000. “Nutrient Transport and Water Quality Monitoring in Sabine Lake Bayous”, Progress reports 1, 2, and 3, and the final report for the Texas Water Development Board.

Gleit A., 1985. “Estimation for small normal data sets with detection limits”. Environmental Science and Technology, 19(12):1201-1206.

Hydrolab, 1998. Hydrolab Multiparameter Water Quality Monitoring Instruments, Operating Manual, Hydrolab Corporation, Austin, Texas.

Longley, W. L. (editor), 1994. Freshwater inflows to Texas bays and estuaries: ecological relationships and methods for determination needs. Texas Water Development Board and Texas Parks and Wildlife Department, Austin, Texas. 386 pp.

Wanielista, M. A., and Yousef Y. A., 1992. Stormwater Management. John Wiley & Sons, Inc. 600pp.

YSI, 1998. YSI 6920 Multi-Parameter Water Quality Monitor, Instruction Manual and Service Manual. YSI Incorporated, Yellow Springs, Ohio.