In this issue:

• Dow Chemical Uses iThink to Find Supply Chain Improvements
• Systems Thinking and Dynamic Modeling Deliver New Insights into Freshwater Systems
• Top Blog Posts of 2009
• Upcoming Web Seminar Series

Martin Fernandes
Dow Chemical
  

Over time, that initial tinkering has led to more than 100 iThink models that helped discover supply chain improvements and address issues faced in Six Sigma recruitment, business strategy, accounts receivable and payable, and health and safety.

Fernandes points out that many modeling exercises are inspired by a seemingly simple question that can be answered using a widely shared mental model. Of course, the worry is that the mental model isn’t exactly right. He points to a supply chain question that manufacturers and engineers needed to answer – how many new chemical storage tanks would be needed to support new products?

By looking at the number of tanks used to support current products, the team calculated they’d need 3 or 4 new tanks to accommodate new products. They asked Fernandes to help them figure out it they were right.

“I encouraged the group to take a step back and look at their manufacturing processes,” says Fernandes. “This seemed like a great chance to combine the concepts of lean manufacturing with Systems Thinking.”

“Whenever I build a model I spend about 80% of the time looking at the problem and gathering data and then 20% of the time actually building and validating the model,” says Fernandes. He gathered information about the plant that would be manufacturing the new products. It was located in China and included the latest technology. Dow also kept records on the material that was in each tank. “We knew what each tank held and how much for every minute.” After spending three weeks collecting data , Fernandes built an iThink model in two days that considered production and decision-making processes, system capacities, and constraints.

Next, he brought together a cross-functional team of US and China-based employees to validate the process he had modeled and then used it to test assumptions about the tanks they’d need to add. The model was linked to an Excel spreadsheet allowing them to easily change system inputs and other variables. Together, they looked for and found leverage points where changes in the system itself could yield better performance.

“The team discovered that it didn’t need to add any new tanks,” says Fernandes. “The Chinese plant already had 12 tanks and by making various improvements, they only needed 10 to support current products. That left two tanks for the new products. The team broke the old mental model that said ‘The tank size is X so we need Y tanks for Z products.’ When everyone worked together, they found a way to change their system and save money.”

That supply chain model can be applied to similar questions at other plants. It has to be tweaked to account for different technology that’s in use or other local variables but the same concepts apply.

While the models that he’s created have helped answer questions and solve problems throughout the company (and in locations around the world), Fernandes is quick to point to the work teams that are involved. “The work teams are the ones providing information for the models and using them to simulate scenarios and find leverage points. iThink is an enabler,” he says. “It allows people to challenge their mental models and test new ideas.”

“iThink is so easy to use,” says Fernandes. “The ability to create an interface with sliders and other graphical devices, lets people focus on their ideas and simulation outcomes, not the mechanics of the software.”

iThink is being used widely throughout Dow Chemical. Fernandes provides training and coaching for new employees and acts as an internal consultant when other iThink users have a question or need advice. “Even our leadership uses iThink to work through scenarios and discuss new ideas. Its always better to test assumptions than to make a guess.”

Stephen Page,
Fisheries & Oceans Canada
 

“We look at the big picture and changes over time,” explains Stephen Page, an Aquatic Chemist at Fisheries and Oceans Canada, a federal department that manages the ELA for its research. “We study all parts of the lake ecosystem: the watershed, the fish, hydrology, phytoplankton, zooplankton, invertebrates, everything.” This approach has allowed Page and other scientists to understand, for example, how acid rain affects fish, whether high phosphorous levels cause eutrophication, how aquaculture (fish farming) in a lake may change the entire system, and how mercury accumulates as it moves from the atmosphere into a lake and through the aquatic food chain.

Recently, Page (and colleagues Hesslein, McCollough, and Stainton) applied that whole-system, big picture approach – a hallmark of Systems Thinking - to study nutrient concentrations and loading in Lake Winnipeg. Over the past 10 – 15 years, the algal blooms on Lake Winnipeg have increased in size, frequency, and intensity; a sign of increased nutrient concentrations and deterioration. Page and his colleagues hoped to effectively assess proposals that had been made to reduce nutrient loading and concentrations and, a result, the algal blooms.

The increase in algal blooms has been a very noticeable and unwelcome change. The blooms ruin the lake’s aesthetic quality, are a nuisance to people who use the lake for recreation, and an inconvenience for fishermen who have to deal with nets covered and clogged with diatoms. Some blue-green algae are dangerously toxic for humans, livestock, and other animals (pets) making it a concern to everyone near the lake. Noticeable, annoying, and scary as they are, the algal blooms have increased environmental awareness among Canadians, Manitobans in particular.

Algae are fed by nutrients (phosphorous, carbon, and nitrogen). Key to finding out why they were increasing was identifying the sources of those nutrients and understanding how they could be reduced. The Lake Winnipeg watershed is very large. In fact, it has the largest land drainage to lake surface area ratio of any of the great lakes of the world. The Red River, Saskatchewan River, and Winnipeg River are the main tributaries that feed the lake. The initial assumption was that an increase in nutrients from lakeside industry, residences, tributaries, and land runoff was fueling the increase in algal blooms.

Given that assumption, several nutrient reduction strategies had been proposed by various regulatory bodies.

• Reduce nitrogen loading by 13% and phosphorus loading by 10%
• Return Lake Winnipeg to a prior condition (i.e. “1970 loading conditions”)
• Remove nitrogen and phosphorus from point sources (i.e. manage urban effluent from the City of Winnipeg to 400 tons)
• Manage the lake to a target phosphorus concentration (eg 25mµg/L TP)

Running experiments to determine the effectiveness of any of these proposals is somewhat impractical, if not impossible. It would certainly be time consuming. Fortunately, hydrometric (flow) data has been collected on and around Lake Winnipeg since the early 1900s. Page and his team were able to create a dynamic model that used the data to describe changes in the lake over the last century and into the future.

The Lake Winnipeg model estimates all sources of nutrient inputs – rivers, run off, direct input from people, rain, etc. Since the team had discharge data for each input and had generated modeled concentrations of nutrients in each input, they were able to estimate the monthly nutrient load to the lake as well as lake nutrient concentrations. Those estimates allowed them to simulate the effectiveness of the four proposals.

Simulations indicated that, among Lake Winnipeg tributaries, an increase in phosphorous-rich water from the Red River was the dominant factor in the increased algal blooms. Perhaps more interesting, and certainly surprising, was the failure of any proposal to significantly reduce the phosphorous concentrations which lead to an increase in the size, frequency, or intensity of the algal blooms.

Why? Rainfall in the region had increased more than 20% over the last decade. More rain water is moving through the watershed and into the lake. Even if, for example, the city of Winnipeg drastically reduced its nutrient contribution, more water was running into the lake from rivers and land areas and that increased flow carried more nutrients. Any manmade reduction efforts would be offset. And, because humans can’t control rain amounts, it would be very difficult to manage the lake to a set level or return the lake to a prior condition.

Page wondered what would happen if the increasing precipitation trend continued or if there was a drought. As already explained, when rainfall is high, increased flow into the lake offsets any reduction in nutrient load. In the case of drought, the load of nutrients flowing into the lake decreases and attempts to reduce nutrient load could reduce the algae population. Of course, with fewer algae the fish would have less food and under very low flow conditions the fisheries stock could potentially decline.

The model and simulation was instrumental in setting people’s expectations for their ability to manage nutrients in Lake Winnipeg. During rainy times, not much can be done. During droughts, it might be possible to find what Page calls a “sweet spot” in which phosphorous concentrations and algal blooms decrease but not to the extent that fisheries also decrease.

“Systems Thinking and the dynamic model really helped us break a mental model and think in a whole-system way,” says Page. “We probably could have done the analysis with a spreadsheet but it would have been ugly and we wouldn’t have had a good way to communicate results to involved politicians and concerned people. I could have built a PowerPoint deck with our findings but the minute someone asked, ‘Yeah but what about X?’ I would have had to answer, ‘I’m not sure. I don’t have a slide for that.”

Using his model, Page was able to present results and engage his audience’s “what if” questions. “With a model and simulation capabilities I can say ‘Let’s try that,” says Page. “People don’t need to be modelers to get insight from a simulation. When they see how outcomes change, they develop confidence in a new model and an ability to think in a systematic way. Everyone gets on the same page.”

Fisheries and Oceans Canada have been applying Systems Thinking creating dynamic models of freshwater systems for years. “It’s great to have a consistent way of working through a problem that can be followed by scientists and laypeople,” says Page. “Systems Thinking and dynamic modeling give us the concepts and tools we need to work through the science, discover new insights, and test solutions.”

One of ELA’s “new, cool” experiments is focused on the impact of water diversion and water reallocation. It seems simple, if water from upstream lakes stops flowing into a lower order lake, it will eventually dry up. That, it seems, doesn’t necessarily happen based on their preliminary simulations. It’s just another incorrect freshwater mental model that Systems Thinking and a dynamic model are helping to break.

As the first anniversary of the isee Blog approaches, we thought it would be interesting for folks to see the list of our most popular blog posts:

1. Modeling H1N1 Flu Outbreak
2. Modeling Customers Switching Between Brands
3. Modeling a Watershed with Arrays
4. Matrix Arithmetic
5. Spatial Modeling with isee Spatial Map
6. “Thinking in Systems” book inspires online course
7. Physics Textbook 2.0
8. Insight-based Model Investigates the Housing Crisis
9. Building a Health Care Model Hierarchically
10. C02 in the Atmosphere Behaves Like a Bathtub

Dynamic Dashboards: Practical Tools for Improving Performance & Communicating Strategy Join us for this two-session live web seminar series Beginning January 8, 2010 11:00 - 12:10 PM EST Presented by: Chris Soderquist President, Pontifex Consulting   To learn more or to register, visit www.iseesystems.com/dashboards