3

Review of Calculator Algorithms and Mechanics

USCG INLAND ERSP CALCULATOR FRAMEWORK AND OPERATIONS

The calculator has two modules. The Oil Behavior Module is used to set the spill scenario. The Recovery Module (not actually identified as such by the developer) describes the recovery system and operating environment and calculates the Estimated Recovery System Potential (ERSP) value.

The Oil Behavior Module relies on a set of lookup tables that were developed based on simulations of the SIMAP model (see USCG, 2021a, app. E). The Oil Behavior Module allows a user to select one of four oil types (Group I, II, III, or IV). There are two Waterway Configurations (open and confined) that can have two Current/Flow regimes (flow and stagnant), thus defining four possible types of waterways. There are two shoreline types (gravel and wetland). The flow regimes determine the default mixture of the two shoreline types. Three volumes of oil can be used (50, 2,500, and 155,000 bbl). With 4 waterway configurations and 3 oil volumes, there are 12 scenarios for each oil type. The calculator provides solutions based on continuous releases (conditions do not change with time) or batch releases (conditions change per release event). Under batch releases, it provides output for 3 days. Conditions over the 3 days are based on the 12 scenarios for each oil type. Under continuous release, the conditions remain constant until the release is stopped.

The waterway configurations used in the SIMAP model to develop the lookup tables used in the calculator include several items. A model of the lower Delaware River was used to represent an open system with flow. A model of a section of the St. Johns River was used to determine conditions in a confined system with flow. The open-stagnant system weathering is based on a 5-nautical mile (NM) by 5-NM grid and confined-stagnant weathering is based on a 5-NM by 1-NM grid. The stagnant systems are assumed to have wetland shorelines. Therefore, a total of 72 model runs comprise the lookup table rather than the 96 that would be needed if all 48 possible scenarios needed both gravel and wetland shorelines.

Customization Module A allows a user to select a different proportion of gravel and wetland shoreline types. The calculator uses a linear interpolation of the two shoreline types for the specified scenario to establish the values passed to the Recovery Module. This expands the possible scenarios from 12 per oil type to over 600 per oil type. In the absence of the sensitivity analysis, it is not possible to evaluate which of the proposed values or scenarios actually add value for the determination of estimated recovery system potential.

The Oil Behavior Module provides a daily oil thickness and percent emulsification to the Recovery Module. The daily oil thickness is the combined thickness of oil and emulsions.

The Recovery Module sets the number of hours the system operates each day based on the Operating Period minus the Downtime. The operating hours on day 1 are also reduced by the Recovery Start time. The Speed of the system, or current strength, defines how fast oil enters the skimming system. The user then provides the Swath Width of the system for each of the three operating days. The amount of oil and emulsion available to the recovery system (Encounter Rate) each hour is determined by the Swath Width times Speed of Advance times Daily Oil Thickness (from Oil Behavior Module). Multiplying the Encounter Rate by the Operating Hours, the calculator determines the amount of oily product that could be recovered each day.

The recovery system is defined by the Swath Width of the system, the Throughput Efficiency, and the skimmer characteristics of Recovery Efficiency and Nameplate Recovery. There are three selectable Nameplate Recovery values (250, 1,000, and 2,500 gallons per minute). Storage is not considered part of the recovery system design and is effectively infinite. There is a Maximum Effective Swath Width (MES) that is determined by when the Encounter Rate times the Throughput Efficiency equals the amount of oil

and water that the skimming system can recover (determined by the Recovery Efficiency and Nameplate Recovery). The calculator determines the MES and will not run if the user entered Swath Width exceeds the MES. Thus, a user must use trial and error to determine the MES.

The calculator, in its standard configuration, provides the daily ERSP and total ERSP as its only outputs. When the system is below the MES, the calculator determines the daily ERSP by the Encounter Rate and the Operating Hours. If the daily ERSP values sum to less than the user-specified Spill Volume, then the total ERSP is the sum of the daily ERSP values. If the sum of daily ERSP values is greater than the user-specified Spill Volume, then the total ERSP is set to equal the user-specified Spill Volume. Note that the overall ERSP value can be less than the individual daily ERSP because the overall ERSP is forced to not exceed the volume of oil spilled, whereas the daily values are based on the assumed recovery from an infinite area slick of a fixed thickness. The daily and total ERSP are only 2 of the 12 items that the calculator computes. Many of the other values calculated determine the water recovered but then are not used because the model does not consider storage. Additional outputs can be obtained by using the Customization Modules B and C (e.g., USCG, 2021c).

Module B provides the recovery at a specified hour. An hourly thickness is used in this module rather than the daily thickness provided in the main calculator. This allows significantly more oil to be collected in a single hour than the calculator estimates for the full day (e.g., hourly recovery can be >5,000 bbl/hr with daily ERSP being 1,500 bbl/day). Since the sum of the hourly ERSP is not equal to the daily ERSP, it leads to confusion about the actual recovery potential of a particular system. The module also provides an estimate of time to no recoverable oil. However, the number of hours to no recoverable oil does not change even when the hour input into the module is varied. The hours to no recoverable oil are the same at hour 6 and hour 10 of recovery. It does not appear that the hour in this module is tied to the operating hours provided in the main calculator.

Module C provides the amount of recoverable oil at a specified hour. Modules C and B cannot be run at the same time. The module provides the Percent Recoverable Oil at the specified hour and at the end of each of the 3 days. It also provides the Volume of Recoverable Oil at the hour and end of the 3 days along with the Time to No Recoverable Oil. Once again, the Time to No Recoverable Oil is not tied to the number of hours that oil has been collected. The Volume of Recoverable Oil can be significantly less than the estimated recovery each hour because the volume uses the user-supplied input, whereas the ERSP is not calculated based on the user volume.

Underlying Assumptions

The committee found that there were several underlying assumptions embedded in the calculator that are important to consider when assessing the applicability of calculations. These assumptions were either found within the documentation or based on calculator operations.

1. Best response practices are used and equipment is positioned to optimize the recovery (e.g., the skimming system can always remain in oil). The way this assumption is implemented is to provide oil throughout the day at the average thickness. This leads to a situation in which the area, and therefore volume, of the slick is infinite. Since the thickness is fixed each day, the volume of oil depends on the specified speed of the current or advancing skimmer. This leads to situations where the ERSP can be several times the spill volume. While the design documentation references calculating the oil volume on the surface (see USCG, 2021a; important for the Percent Recoverable Oil output of the Oil Behavior Module), there is no evidence that the actual volume of oil provided by the scenario as defined by the user inputs is calculated.

   There are times that the assumption that the skimmer remains in oil is reasonable and it is consistent with the assumption used in the BSEE Offshore Calculator. During a continuous spill, one can expect that there will always be oil flowing to a skimming system. In a batch spill, using open-water recovery, it can be assumed that the skimmer is continually collecting oil until it collects all

of the recoverable oil. This assumption may not be valid for shoreside recovery of smaller spills that could pass the recovery system before the end of the 3-day period. With information about the width of the waterway and the waterway speed, it would be possible to determine if a shoreside recovery system could remain in oil through its operating period.

1. The recovery does not affect the daily thickness. This assumption is reasonable if the amount of oil recovered is a small percentage of the recoverable oil. It can be a reasonable assumption until all recoverable oil has been collected, but at that point the oil thickness should reflect that there is no available oil. If available oil remains, this assumption requires that the spill has infinite volume.
2. There is infinite storage available. Ignoring storage is problematic in that storage is directly related to how EDRC (which drives the current classification system) is calculated, and how the existing OSRO classification system is based. The assumption of infinite storage is not consistent with ASTM Standard F1780-18 on calculating ERSP. In remote locations, storage can be the limiting factor. Even in more accessible locations, the time required to change storage units is important to understanding a system’s potential performance. The lack of outputs other than daily ERSP does not allow the user to estimate how often storage units would need to be changed. Understanding the role of storage in a recovery operation can also be a valuable learning opportunity for inexperienced planners.
3. Skimmer nameplate capacity can be simplified to three values. The calculator allows the user to select one of three values for the Nameplate Capacity (250 gpm, 1,000 gpm, or 2,500 gpm). A quick check of the smaller skimmers that are likely to be used in shallow-water environments showed that the lowest available capacity in the calculator was still a factor of 10 greater than the numbers provided by the manufacturer.¹ This leads to maximum effective swath widths and ERSP values that are highly inaccurate and misleading. This input is important enough to the calculations that it is poor to assume that it can be simplified in this manner.
4. The oil is spilled into the waterway and later reaches the shore. This assumption may result in more oil being available for on-water recovery than appropriate. For the purpose of the calculator, this may be a reasonable assumption, but it should be clear in the documentation.
5. Each of the waterway configurations can be modeled based on a single waterway configuration. For example, all open-flowing rivers are equal to the one section of river that was selected to model. The expectation is that there are numerous aspects of the waterway that will affect the oil thickness and weathering processes. There is no evidence that a sensitivity analysis was completed to determine the most important environmental and waterway characteristics that must be accounted for when estimating the oil thickness. Customization Module A allows a user to vary the shoreline type. However, it is unclear why the shoreline type should be expected to be more important than the waterway width in determining the oil retained on the shoreline. This is the step that begins to include some of the complexity of the inland environment, and it is difficult to know when to stop adding levels of complexity. Thus, it may be reasonable to only use one potential scenario for each of the waterway configurations. It would be beneficial to conduct a sensitivity analysis to get a better understanding of how representative the results may be and guide future calculator modifications.

Differences Between How the Calculator Operates and How It Appears to Operate

There are several substantial differences between how the calculator actually operates and how the user may assume it operates based on the interface. These differences occur when inputs are not used or are not accurate reflections of the information being used.

___________________

¹ For example, it is possible to select a Minimax skimmer, which the manufacturer lists at 20 gpm recovery. However, the minimum recovery considered in the calculator is 250 gpm and the calculator allows the skimmer to be given a recovery rate of 2500 gpm. See https://www.elastec.com/products/oil-spill-skimmers/drum-oil-skimmers/minimax/#specs, accessed May 11, 2022.

Many inputs appear to only be available to help a user identify the inputs they need to enter. These include Planning Volume Classification, Latitude Zone and Season, and Source Type. It is not apparent to the user if these inputs are providing suggestions or populating a field in the calculator, which creates confusion about their purpose.

The Recovery Type entry appears to allow a user to select between open-water and shoreside recovery systems, but the underlying calculations, particularly encounter rate, are identical for both types of systems so there are no differences in the ERSP. This is problematic when considering shoreside recovery systems in which the recovery should reflect the waterway conditions if a user wants to know how much oil passes out of the reach of the swath width or how long oil can be collected before the spill moves past the system.

Spill Volume is a deceptive input. A user appears to be able to enter any spill volume desired with the expectation that the system performance will be associated with that volume. However, there are three spill volumes, and the selection of which of the three volumes will be used depends on the spill volume entered (USCG, 2021a). For spill volumes of 1 to 999 bbl, a spill of 50 bbl is used. For a spill volume of 1,000 to 99,999 bbl, a spill volume of 2,500 bbl is used. For spill volumes of 100,000 bbl and greater, a spill volume of 155,000 bbl is used. Although the Spill Volume input appears continuous to a user, it is effectively three possible volumes.

Similarly, despite an extensive library of skimmer choices that makes it seem like the user can select the characteristics of very specific skimmers, the nameplate capacity of the skimmer is not linked to the skimmer. Instead, the software uses a Nameplate Recovery range selected by the user. The Nameplate Recovery range appears to be valid for particular ranges of values; however, only three specific values (250, 1,000, and 2,500 gpm) are used (USCG, 2021a).

The automatic selection of a Throughput Efficiency with the selection of a skimming system is questionable. The Throughput Efficiency in inland response configurations is highly dependent on the boom being used, how it is deployed, and the speed of the water/system. It is not a factor that should be attributed to a skimmer.

Functional Differences Between the BSEE ERSP Calculator and the USCG Inland ERSP Calculator: What Are They?

One charge to the committee was to evaluate consistency with the BSEE ERSP Calculator. The BSEE ERSP Calculator provides only one set of daily oil thickness and emulsion percentages. The selection is based on either a batch release, in which the variables change per release event, or a continuous release, which repeats the day 1 conditions. In contrast, the USCG Inland ERSP Calculator has four oil types and three oil volumes, two waterway configurations, and two flow regimes with associated shoreline types (ignoring Customization Module A). Thus, the USCG Inland ERSP Calculator has 48 basic scenarios that can be selected. This means that for one skimming system (i.e., boom and skimmer) the user will have 48 inland ERSP values and 96 inland ERSP values if they carry out batch and continuous releases for each of the 48 scenarios.

The BSEE ERSP Calculator considers storage and the USCG Inland ERSP Calculator does not. The BSEE ERSP Calculator uses information about the storage size and downtime associated with changing storage to calculate the amount of recovery system downtime. Users must reduce the number of operating hours if they feel that the system will be down for other factors. The USCG Inland ERSP Calculator requires users to enter a downtime value that covers the time associated with changing out storage and any other reason. Since the downtime value in the USCG Inland ERSP Calculator does not change daily, there is no way to account for changes in the storage downtime with changing oil thickness. The lack of oil and water recovery outputs makes it extremely difficult for users to determine the potential storage requirements and downtime using the USCG Inland ERSP Calculator.

The BSEE ERSP Calculator allows for decanting and the USCG Inland ERSP Calculator does not. Decanting is only important if storage is being considered and a user is trying to maximize the amount of oil in storage.

The USCG Inland ERSP Calculator only uses three Nameplate Recovery rates, whereas the BSEE ERSP Calculator allows any value. The Nameplate Recovery rate is important in determining the MES and ERSP. By limiting the possible Nameplate Recovery rates, the inland ERSP severely restricts the potential recovery systems that can be examined using the USCG Inland ERSP Calculator.

The BSEE ERSP Calculator determines the MES and automatically reduces the Swath Width used in the calculations to the MES if the user-provided value is too large. The USCG Inland ERSP Calculator requires the user to use trial and error to find a swath width below the MES.

The BSEE ERSP Calculator uses full recovery days in each of the 3 days, whereas the USCG Inland ERSP Calculator reduces the recovery on day 1 based on the time to deploy equipment.

The BSEE ERSP Calculator provides detailed outputs on the conditions being considered (Oil Thickness and Percent Emulsification), oil and water recovery through time, time recovering oil, and downtime associated with changing storage. There is detailed information provided about the encounter rate, the recovery rate of oil and water, storage variables, and the volume of oil and water recovered. The result is the daily ERSP. The USCG Inland ERSP Calculator only provides the daily ERSP and no information about the conditions being considered, recovery rates, or recovery volumes. Without knowing fluid recovery rates, a user cannot estimate the amount of downtime needed to change storage. The lack of outputs also makes the tool less appropriate for understanding how to optimize recovery systems.

Functional Differences Between the BSEE ERSP and the USCG Inland ERSP Calculators: Do They Matter?

The primary addition in the USCG Inland ERSP Calculator is the Oil Behavior Module which provides a wide range of potential scenarios. The lack of scenarios in the BSEE ERSP Calculator makes it easier to use for planning purposes (i.e., farther to the left of the Figure 2.1 continuum) because it is simple, and it is easy for the planner and anyone evaluating compliance to arrive at the same solution. While the BSEE ERSP Calculator only considers one scenario, since the encounter rate is a linear function of oil thickness (e.g., decrease the oil thickness by half and the recovery rate decreases by half), it can be used to estimate recovery in a much broader range of scenarios. This is true as long as the system is operating below the MES. This is important because, even though the BSEE ERSP Calculator provides only one scenario with a given daily oil thickness and emulsion percentage, a user can quickly estimate the ERSP of thinner oil thicknesses with that emulsion percentage. Since the encounter rate is a linear function of speed, swath width, and oil thickness, it is possible to quickly determine changes in how swath width or speed can vary with oil thickness.

The USCG Inland ERSP Calculator ignores storage and limits the Nameplate Recovery values. Ignoring storage is problematic in that storage is directly related to how EDRC is calculated and how the existing OSRO classification system is based. Furthermore, if storage is ignored and a user is not concerned with knowing how much water is collected, the emulsion percentage is less relevant and the input needed from the spill scenario is reduced to only requiring the oil and emulsion thickness. The BSEE ERSP Calculator then can be used to quickly give an ERSP for any oil and emulsion thickness (below MES). Since storage is not considered in the USCG Inland ERSP Calculator, the BSEE ERSP Calculator can be used to provide an ERSP for all the scenarios included in the USCG Inland ERSP Calculator and more.

The Recovery Module of the USCG Inland ERSP Calculator only allows three possible Nameplate Recovery rates, which means that very few potential recovery systems can be accurately entered into the calculator. The performance of many, if not most, of the recovery systems likely to be used during a spill response cannot be estimated using the USCG Inland ERSP Calculator.

The lack of outputs in the USCG Inland ERSP Calculator limits a user’s ability to understand how to optimize a recovery system. An important estimate that the USCG Inland ERSP Calculator could provide is the starting oil thicknesses for the different oil types and system configurations, but that is not an output provided by the calculator. One can look up the daily oil thickness values provided in the appendixes of the USCG Inland ERSP Calculator design document and then apply them to the results of the BSEE ERSP Calculator. Without recovery rates it is not possible for a user to estimate the amount of downtime associated with changing storage.

VALIDATION AND VERIFICATION

To test the inland ERSP calculation, the USCG contractors performed a verification and validation study for the calculator (USCG, 2021c). The purpose of the verification (code) study was to verify the functionality of the Prototype Production Ready Stand-Alone Inland ERSP Operational Environment Calculator Tool and make sure it was consistent with the design, as outlined in the Design Document (USCG, 2021a). The goal of the validation portion of the study was to determine whether the results from the calculator tool are realistic.

In the verification process, more than 20 cases were explored, varying the input to the calculator and the customization modules (see USCG, 2021c, tables 4–6). The cases tested did not cover the full range of the potential inputs. In the verification process, six cases were explored varying the oil type, spill volume, swath width, shoreline type, recovery at specified hour (Customization Module B) with varying shoreline (Percent Wetland in Customization Module A), and identifying recoverable oil (Percent and Volume) at specified hour (Customization Module C) with varying shoreline (Percent Wetland in Customization Module A). The metric for the validation study was whether the developers of the USCG Inland ERSP Calculator thought the results were reasonable and internally consistent with other cases that they had simulated.

Several of the committee members ran hundreds of simulations to understand how the calculator worked. One problem that immediately occurred was that the calculator was programmed using the Windows 10 operating system and Microsoft Office software, using the Visual Studio 2019 IDE, and written in C#.net to produce an installable Windows Form Application. This presented a challenge for those using Apple systems.

The committee generally was able to reproduce the test cases provided in the verification and validation report. In exploring other cases, the testing identified problems with the implementation and several areas that lacked internal consistency. Some of these issues are documented in Appendix B. However, Appendix B should not be considered an exhaustive listing of potential problems with the existing calculator. It merely lists some of the issues identified during the committee’s testing. Although it is not possible to identify the source of all of the problems, several issues were revealed. It appears that there are not consistent checks to make sure that the values selected as input are reasonably constrained, and the verification process did not cover the full range of situations allowed by the calculator. For example, the previous verification process was restricted to low-speed recovery, whereas many problems became more evident when looking at high-speed recovery systems. Some of the inputs allowed were also unrealistic.

In development of the Inland ERSP model, there have been no constraints imposed in the modeling framework to ensure that the model maintains conservation of oil mass throughout the simulation. Given this situation, it is possible for the model to predict more oil recovered than was spilled. For example, this routinely occurs if the forward speed of the skimmer or the stream-flow speed is selected to be a large value. Values as large as 100 knots can be selected by the user. The USCG Inland ERSP Design Document indicates a default value of 0.75 knot (User Defined Input UDI-8; USCG, 2021a) with a note that fast booming is required if the value exceeds 0.75 knot. It is noted in Appendix F, Figure F-1 of the design document that the speed can increase as the boom angle is reduced, reaching a value as large as 6 knots if the boom angle is reduced to 10 degrees (USCG, 2021a). A review of detailed numerical modeling of the performance of oil booms, with verification by laboratory experiments conducted at the Ohmsett (the National Oil Spill Response Research & Renewable Energy Test Facility) and University New Hampshire facilities and

funded by U.S. Department of Transportation and Minerals Management Service by Grilli et al. (1996a,b; 1997; 2000), Hu and Grilli (1997), and Grilli and Hu (1998) give consistent results with peak flow speeds of 0.75 to 1 knot before boom failure. The failure speed is also dependent on the relative density between the spill and the receiving water, and the surface tension of the oil–water interface. These can be accommodated but require a more sophisticated modeling framework such as SLICKMAP (Grilli et al., 2000). Based on this analysis, boom failure could be mitigated and the calculator could remain simple to use by limiting USCG Inland ERSP Calculator speeds to no more than 1 knot with regular booming and 6 knots for angle booming.