Duy trì chất lượng nước của Nam Florida với Minitab

02/07/2019 846 lượt xem

Coastal and river waters are critical to Florida’s economy, providing jobs via recreation, fishing, and tourism. Protecting those waters helps safeguard Florida’s economic health, but effective environmental management decisions need to be based on sound research. The Southeast Environmental Research Center (SERC), located on the main campus of Florida International University, unites researchers from a variety of disciplines to conduct scientific investigations in threatened environments of South Florida and the southern United States. The center’s research projects in Biscayne National Park, Big Cypress National Preserve, Everglades National Park, Florida Bay, the Florida Keys and the Florida Keys National Marine Sanctuary have provided a basis for management decisions for sustaining these fragile resources. Combining data from all of the center’s projects gave the researchers an opportunity to develop a unique, high-level view of conditions throughout South Florida waterways—but joining and analyzing 20 years worth of water quality data collected from hundreds of water monitoring stations was a major challenge. They needed to understand what their enormous wealth of data revealed about the health of Florida’s waters. For that, SERC’s researchers trusted Minitab Statistical Software.

Creating this overview of Florida’s waterways was part of a National Park Service initiative to support the Florida Department of Environmental Protection and the US Environmental Protection Agency to develop numeric water quality criteria for the protection of aquatic life and human health. Years of data monitoring had already provided some insights into the relationships between water nutrient levels and pollution—for example, increased levels of nutrients, such as phosphorous and nitrogen, produce rapid algae growth. Rapid algae growth makes waters cloudy and depletes oxygen from the water column, causing fish populations to die-off. SERC researchers wanted to investigate further and find exactly what nutrient levels would trigger unwanted algal blooms in South Florida waters.

A Minitab Censored Probability Plot of the pollutant NOx present in the Central Florida Bay made it easy for SERC researchers to see how contamination levels changed over time.

Dr. Henry Briceño, FIU research associate and Principal Investigator at SERC, led the project. He and his team faced the task of analyzing literally millions of data points with samples collected from 350 stations across South Florida.

Water quality parameters monitored at each station included dissolved nutrients such as nitrite, nitrate, ammonium, inorganic nitrogen, and reactive phosphorus. Total concentrations of nitrogen, organic nitrogen, phosphorus, and organic carbon were also measured, as well as chlorophyll-a, alkaline phosphatase activity, salinity, dissolved oxygen, temperature, pH and turbidity. All parameters were measured at both the water surface and bottom of the water column.

A Minitab Censored Boxplot of the pollutant NOx present in the Central Florida Bay made it easy for SERC researchers to see how contamination levels changed over time.

Some of these observations were found to be "nondetects" —meaning the contaminants in these samples fell below the reporting limits of the measurement instruments. Nondetects are important to consider because even low-level contaminants can prove to be an integral part of the data as a whole. This posed a challenge for the researchers, because it wasn’t clear how to account for nondetects using standard statistical procedures, particularly given the huge quantity of data involved.

To help, SERC enlisted the expertise of Dr. Dennis Helsel, owner and lead scientist of Practical Stats (PracticalStats.com), an environmental statistics training and consulting firm. Dr. Helsel specializes in statistics for data with low-level contaminants. For this project, he utilized Minitab Statistical Software and its powerful Macros feature to efficiently handle statistical calculations on the datasets containing nondetects.

A Minitab Macro consists of a set of commands stored in a text-based file that can be tailored to meet an analyst’s specific needs. With a single click, each command within the macro is executed instantaneously and the analysis is complete. Macros save time, and are frequently used to automate repetitive tasks and quickly perform extensive data manipulation. And for specialized tasks such as this one, macros can also expand Minitab functionality by computing custom statistical procedures. Using methods he outlined in his textbook Nondetects and Data Analysis, Dr. Helsel wrote macros that adapted Minitab’s censored data techniques—those techniques that help to account for missing or non-detectable data—to work with SERC water quality data. This made the vast quantity of water quality data much easier for SERC researchers to derive the required statistics for data analysis.

Minitab Multivariate Analysis helped researchers reduce large amounts of water quality data into more manageable variables for classification. SERC researchers were able to group monitoring stations into geographical zones by water type, pictured above.

Dr. Helsel trains scientists like Dr. Briceño to perform statistics during their research and he applauds Minitab for its "ease of use." "Minitab is very easy to teach with," Helsel says, "Scientists are able to learn and use the software immediately."

Dr. Briceño is not a statistician by trade, and he easily used Dr. Helsel’s macros and Minitab to calculate the statistics necessary to analyze water quality at each monitoring station. Monitoring stations with similar statistical outcomes were clustered into groups sharing common characteristics with a combination of Minitab Multivariate Analysis tools. Mutlivariate analysis is used to reduce large data sets into a smaller number of components, making it more efficient for researchers to hone in on specific variables. SERC was then able to group monitoring stations into different "water types" and make detailed maps of their geographical distribution (Fig 2). With the water bodies classified, the concentration levels, or thresholds, of phosphorous and nitrogen that would cause deadly algal blooms to thrive were assessed for each water type. This was achieved by plotting cumulative sum charts of chlorophyll-a along nutrient gradients in Minitab. These charts made it easy for researchers to quickly read the threshold levels directly without further analysis.

Minitab Cumulative Sum charts helped SERC researchers easily identify threshold levels of phosphorus and nitrogen that cause deadly algal blooms to grow.

SERC researchers were able to formulate innovative methodologies to derive protective nutrient criteria for each individual water body. The knowledge gained from FIU and SERC studies and their Minitab analysis will help scientists plan for continual conservation and restoration. Best of all, the information learned in South Florida can be beneficial for scientists working to correct pollution all over the world. These methodologies are currently under consideration for widespread use by the United States Environmental Protection Agency.

Dr. Briceño says this project has shown that South Florida water quality is excellent, and much of that has to do with a consistent focus on research to protect these critical natural resources. And when it comes to the data collection and analysis that underlies that research, "Minitab has provided us priceless tools to improve our craft," Briceño says.

Dr. Helsel’s macros are freely available online at www.practicalstats.com/nada.