This article was originally written and published on a previous version of rasstech.net on 8th of June 2021.

What?

An offshore geotechnical site investigation is a geological survey that employs testing methods from a drilling ship or vessel outfitted with geotechnical equipment and potentially laboratory facilities. An investigation is always commissioned before the construction of new offshore installations but can also occur to measure a change in previously investigated ground. With the advent of larger wind turbines, wind farms are becoming more profitable, and geotechnical investigations are in demand. This article will address geotechnical investigations in the context of the entire offshore energy industry and highlight all the common techniques used prior to construction. Below you can find who, when, where, why, and how of offshore geotechnical investigations. The main goal of this article is to explain offshore operations to interested parties without boring them to death. The secondary goal of this investigation is to assess what trends are being followed and why.

Figure 1: The Kommandor Susan, an offshore supply vessel with geotechnical derrick and drilling setup installed on the main deck. Image used courtesy of Horizon Gardline, and Hays Shipping.

Why?

We require investigations before construction of offshore installations to determine what best depth to lay the foundations, and if it is safe to set up a jack-up. If we ensure the foundations are deep enough, the structure will not tip over. For pipelines and power cables, the main need is to bury the line deep enough to protect it from being uncovered by tidal scour, erosion, fisherman’s nets, current forces, and other passing debris.

Figure 2: A jack-up rig, and offshore supply vessel conducting work on an offshore wind turbine.

Where?

Offshore geotechnical site investigations are done in preparation for construction. The two largest industries for site investigations are oil/gas and wind energy. While other industries exist, this article will focus on wind farms and oil rigs. Wind farms are usually limited to the continental shelf at water depths of around 45m or less, but that is changing with the advent of better foundations. Floating turbines do exist but are still in early development. Wind farms require turbine locations, power station platforms and potentially cable corridors investigated. Oil rig investigations are for platforms, jack-ups, and pipelines. Jack-ups are limited to around 60m water depth, while platforms can reach as deep as 450m. Flowlines and pipelines are commonly laid between platforms and shore stations. Dynamically positioned drilling vessels and floating wind farms require little geotechnical investigation.

Figure 3 (Left): A plan view of a wind farm with cable routing, note the power substation and power convertor. Courtesy of Nordsee One: www.nordseeone.com Figure 4 (Right): A simplified topographic view of oil platforms and subsequent pipelines in the North Sea. Courtesy of: www.petroleumreports.com

Renewables

For most wind farms, investigations are separated into structural and power cable corridors. Boreholes are typically drilled at each turbine location, as well as for the power substation. A small jack-up is used to build and maintain wind turbines and so the investigation must accommodate this. A secondary shallow investigation may be conducted all along the main power cable corridor if it is to be laid below the seabed. A typical investigation for a turbine or substation location is typically 50m below the seafloor, while a cable investigation is usually 6m below the seafloor.

Figure 5 (Left): A wind farm jack-up with defective turbine blade on board, returning after conducting maintenance. Figure 6 (Right): A power station (right) and signal converter (left) located in a wind farm offshore Germany. Cables are routed from each turbine to this power station which aligns the turbine electrical signals and then routes it to shore via the convertor.

Non-Renewables

Oil platforms and drilling jack-ups can be much heavier and exist in much deeper water, so they commonly have stricter geotechnical requirements. Jack-ups usually require an additional Leg Penetration Analysis (LPA) while larger concrete structures like tanks and gravity-based structures (GBS) require much deeper investigations. In the case of Hibernia, a GBS located offshore Newfoundland, geotechnical investigations up to 250m below the seafloor were conducted prior. Pipelines are also laid on or just below the seafloor, which may require shallow geotechnical investigations like a subsea power cable.

Figure 7 (Left): Gravity Based Structure (GBS) Hibernia located offshore Newfoundland and Labrador, Canada. A 250m deep Geotechnical investigation was conducted in preparation. Image used courtesy of Heritage Newfoundland and Labrador www.heritage.nf.ca. Figure 8 (Right): An old 900,000 tonnes concrete protective barrier for a tank located in the Ekofisk field, North Sea. This was the location for the first offshore GBS which has since been dismantled. Figure 9 (Bottom): A large oil and gas drilling jack-up undergoing maintenance and refit in port.

Offshore Geology

Most offshore geology is primarily a variation of silts, sands, and clays which are deposited in the water column over time. Shells, organic matter, larger grades of gravel, and other secondary constituents are commonly deposited alongside sand, silt, and clay. Hard, non-sedimentary rock is encountered closer to shore and must be tested using different techniques than most offshore sediments. There are several methods used for determining offshore stratigraphic profile, these are determined by the construction’s pre-mobilization requirements. Power cable and pipeline investigations are fundamentally different than structural investigations, so the testing methods reflect this. We are more likely to see faster shallow testing methods for cable corridors and deep testing for structures. Oil and Gas investigations are typically more varied in that each structure is a unique rig, whereas wind farms are many similar structures with occasional power stations. Typically, offshore oil and gas structures are much heavier and carry precious cargo in both human lives and environmentally hazardous materials. These risks are considered when crafting a proper geotechnical investigation.

Figure 10: Various seashells found in sample cuttings offshore.

How?

The two main methods used in offshore geotechnical site investigations are sample collection and cone penetration testing (CPT). Sampling is when an in-situ sample of the geology is taken from its natural resting place to be further examined in a controlled environment. Laboratory testing is used to identify the exact soil composition and its engineering properties. CPT testing does not provide a sample but does provide an accurate profile of valuable engineering properties. Both these methods are critical when designing the foundations for offshore structures. There are other methods like downhole seismic systems, which are used mainly to determine layer changes in the geology, and strain moduli.

Sampling

Sampling is when we acquire sections of geology from its natural resting place using a variety of methods. The primary method is using a wireline downhole push sampler inside a drill pipe. Other common methods include using a vibrocore, bucket sampler, or drop sampler. These other methods are only useful for shallow applications, whereas the downhole sampler can be used if the ground can be drilled.

Figure 11 (Left): Shelby tube used to extract samples. A drilling team will choose what type of Shelby depending on ground conditions. These are commonly damaged beyond use when encountering gravel layers. Note the chamfered leading edge, this helps lower the force needed to push the Shelby through the ground. Figure 12 (Right): A real-life soil sample that was extruded from the Shelby tube. In this case, it is primarily very fine sandy SILT, slowly transitioning to clay near the top of the sample. Telling the exact difference between silt and clay in these cases often requires scheduled onshore Particle Size Distribution tests.

Downhole Sampler

Figure 13: Cross-section of a geotechnical vessel undergoing a down whole site investigation. A vessel maintains position over a drill hole using Dynamically Positioning or with a combination of anchors and thrusters. The downhole sampler is the most versatile form of sampling, it consists of sending a Shelby tube on a specialized pushing tool down the centre of the drill pipe and collecting a cylindrically shaped sample of geology from just underneath the bottom of the drill pipe. This method can be used in both shallow and deep applications, and it is the main method used for creating offshore geological borehole plots.

Geo-boring

Geo-boring occurs while drilling into hard rock, typically nearshore. It is like downhole wireline sampling and occurs at the bottom of the drill pipe. Once the drill reaches test depth, the coring tool is dropped into the drill string where it latches near the bottom. The driller will drill the length of the coring tool, filling a liner as it drills. When the liner is full, the bottom will be cut off using a special tool called a cutting shoe. The sample will be de-latched and retrieved to the deck using a special wireline retrieving tool. Like sampling, the rock sample will be visually identified, and tested offshore, or scheduled for onshore testing.

Vibrocore

Figure 14: A vibrocore is used to extract samples up to ~6m below the seafloor. Vibrocore sampling is a form of shallow sampling using vibrations to submerge a long Shelby-like sampling tube into the seafloor. This form of sampling is best utilized for power cable or pipeline corridors as it is much faster to employ than a downhole sampler. It requires a crane or a common A-frame style mounting system rather than a specialized drilling derrick mandatory for downhole sampling. It is not uncommon for geophysical survey vessels to have vibrocore capabilities for shallow geotechnical investigations. A vibrocore will typically use a rotating offset weight to create vibration, like a video game controller or other vibrating adult toys.

Drop/Gravity Sampler

Figure 15: a drop sampler, this sampling method only acquires shallow samples. Is dropped from a height and uses its weight under gravity to penetrate the seafloor. A drop sampler is used in much the same way that a vibrocore is used but requires even less equipment to operate. It is employed from any simple crane on a ship, submerged in the water, and dropped from a height. The advantage of this method is that it requires a very small amount of storage space and is very easy to deploy. The main disadvantage to this method is the quality and length of the sample. Drop samplers also have a difficult time penetrating hard sands and gravel so they also have limited applications to soft cohesive seafloors.

Bucket sampler

Figure 16: The bucket or grab sampling method is comprised of any number of different types of devices that are hoisted to the seafloor and collect a very large undisturbed sample of the earth. This method is not very effective for power cable or pipeline applications and is used more commonly in very shallow seafloor investigations. The big advantage to bucket sampling is that it captures a large area of undisturbed seafloor and so it can be better used to study near-surface geology and used to identify near-surface organic life. It is the best sampling method for keeping organic material intact as other sampling methods tend to crush or slice through shells and organic material. The Sheep’s Foot design you see above belongs to RASS Technical Ltd. Any use of this design must have permission from RASS Technical.

Cone Penetration Testing

CPT testing is the method of pushing a pointy tool into the ground to determine the engineering characteristics of the soil. The CPT tool outputs the pore pressure, tip resistance, and sleeve resistance of the earth as it penetrates. These three outputs can be used to determine the composition of geology. Other useful outputs from the CPT are the inclination, friction ratio, total load force, and speed as the cone penetrates. Understanding CPT data is desirable, and surprisingly uncommon offshore amongst offshore client reps, an experienced eye can usually identify exactly what is happening during the operation, mitigating unwanted time delays. This method is not usually taught in geology programs but is very common in geotechnical engineering. A CPT is normally executed using a downhole wireline tool in the drill string, or with dedicated equipment called a seabed CPT. The seabed CPT tool also has the capabilities to incorporating a seismic receiver, which can be used to gather stratigraphic layer data.

Figure 17 (Left): a CPT cone setup where the tip, pore, and sleeve sensors are indicated. The tip measures the direct force on the cone as it penetrates, the sleeve measures the frictional forces as it passes by, and the pore sensor measures the water pressure as the cone pushes through the ground. Figure 18 (Right): The tip usually reads high in sandy materials, while low in clays/silts, the sleeve reads low in sands while high in clays/silts, the pore pressure reads near static in sands, above static in clays, below static in silts. The “Friction Ratio” is used commonly instead of sleeve friction as it is faster for interpretation.

Downhole CPT

Downhole CPT is the conventional process of testing just below the bottom of the drill string. This commonly occurs offshore for site investigations as it is currently the only reliable way to get CPT data profiles needed for jack-up and wind farm production (40m-50m). The process starts when the driller reaches testing depth, stops rotating, and clamps the pipe with the seabed frame. The CPT tool is lowered down the pipe and latched in place. The test starts and the CPT tool uses hydraulic pressure to propel the CPT rod and cone into the undisturbed ground. Data is collected, the tool is unlatched and lifted to the deck. The driller unclamps the pipe and starts drilling out to the next testing depth. Downhole CPT’s can create a large borehole log, only limited by the ground conditions and the amount of drill pipe on the ship.

Figure 19: A downhole CPT tool is lowered into the pipeline via umbilical and latched into position with CPT cone ready to test in the undisturbed ground just under the drill bit. The test is started, and the downhole CPT tool plunges the cone into the ground. The above figure shows the CPT just as it has begun to penetrate the ground.

Seabed CPT

Unlike Downhole CPT, Seabed CPT does not require a drilling derrick or drilling crew. It can be launched from an A-frame type hoist on all survey ships. The device itself comprises a seabed frame containing a push mechanism and either pre-coiled CPT rods or rods that are assembled onto the string while the frame is lowered. The seabed CPT’s max penetration depth is limited by the strength of the soil and the length of the available CPT rod. For sands will give penetration less than 20 meters while soft cohesive clays can achieve penetration over 40 meters. The value of using a seabed CPT is that you can get deep CPT penetrations without having to conduct downhole drilling, which is both time-consuming and expensive. The secondary benefit is that seabed CPT will produce desirable continuous data coverage as compared to data gaps between downhole CPT’s. Ideally, it is better to use a seabed CPT, but it can be less reliable and is limited by a maximum depth, whereas downhole CPT is not. Seabed CPT is most effective in pipeline and cable corridor ground investigations, where the testing depth required is reliably achieved in all soils.

Figure 20: A coiled rod seabed CPT setup. There are many different versions of Seabed CPT and usually depends on the contractor. Some can be lowered by the derrick crane through the moon pool or others using an A-frame crane.

Sampling Vs. CPT testing

A sampling and CPT log are essential in any geotechnical investigation. Sampling provides the client with an understanding of exactly what type of geology is present, while the CPT log shows engineering properties and highlights layer changes. Typically speaking, a sampling borehole is much more expensive because of additional lab testing and requires more time to complete the same length borehole. This additional cost comes with benefits, as nothing compares with laboratory testing for understanding the ground. The relative densities provided by CPT boreholes must be corroborated with sampling laboratory data to be accurate. CPT data also cannot determine the presence of organics, carbonate content, and other secondary constituents. These secondary constituents, particularly carbonates, can pose a serious risk to future structural integrity. Lastly, not all soils behave characteristically on CPT data and can be misclassified.

Figure 21: Completed preliminary Sampling borehole log and CPT borehole log side by side from the same location (separated by 10m). Note the layer changes are indicated by the same colours. The Offshore Client Rep needs to understand how to interpret and QC these logs as this is the most important deliverable, and what the downstream client pays for.

Hybrid sampling

Hybrid sampling is the process of alternating between sampling and CPT testing in a single downhole borehole. It is a valuable way of acquiring both engineering data, and samples for laboratory testing. A hybrid borehole is quicker than a standard sampling borehole and benefits in the added understanding of density profiles that CPT brings. Hybrid boreholes are valuable to clients who are on a tight budget. The downside to hybrid boreholes is that it requires more experienced engineering judgement as there are pitfalls to the frequency of when to take a sample. It also does not give a continuous profile of sampling data or CPT data, which is useful for applications such as Leg Penetration Analysis (LPA). The last risk of a hybrid borehole is that you risk losing samples of thin layers whilst conducting CPT operations. The most valuable aspect of hybrid boreholes is the ability to use a CPT to power through layers of very dense sands, where sampling is tedious, and relatively fruitless compared to CPT data.

Seismic methods

Seismic methods are commonly used in geophysics to determine stratigraphic layer horizons. They also serve a valuable purpose in geotechnical engineering, especially for structures that have mechanically induced vibrations such as wind turbines. Essentially, there are two main methods, PS logging, and Seismic Seabed CPT. These methods send and receive acoustic signals through the ground to determine the sound velocity profiles of the borehole. This velocity profile is compared with the known density of the previously tested ground to determine the small strain shear modulus, bulk modulus, compressibility, and Poisson’s ratio. These factors are important for assessing risks such as soil liquefaction induced from environmental or mechanically vibration.

Figure 22: Uphole PS logging and Seismic Seabed CPT side by side comparison. A notable difference is the depth of effectiveness. As depth increases, PS Logging becomes more accurate, and Seismic Seabed CPT becomes less accurate.

PS Logging

“PS logging” is the process of acquiring seismic velocity profiles in uncased boreholes. Upon completion of a borehole when the drill is at max depth, a PS logging tool is lowered into the drill string to rest just below the drill bit. A seismic source on the lower part of the tool will sound off, sending acoustic pressure waves into the ground, up the borehole wall, and back into two offset receivers on the upper part of the tool. The offset receivers measure the speed of both the shear wave (S) and the pressure wave (P) as they travel through the geology. This test is repeated at fixed intervals for the length of the borehole. Once acquired, the borehole sound wave profile is coordinated with the density of the previously sampled ground and delivered to the client. In the end, a client gets a density velocity profile of the borehole similar to the seismic SCPT. PS logging results are very noisy near the mudline but have increasingly better resolution deeper in the borehole (greater than 20m).

Figure 23: This is a PS Logger®, where the brown flexible rubber tubing acts as vibrational dampeners between components. There are two hydrophones located at the receivers, and a small acoustic hammer at the source. The PS Logger® is a registered Trade Mark of Robertson Geo Ltd. This image was used with the expressed permission of Robertson Geo Ltd. Shout out to Bobby from Robertson Geo Ltd.

Seismic Seabed CPT

Seismic CPT can be implemented while conducting a seabed CPT investigation. An acoustic receiver is added near the tip of the CPT cone while the acoustic source is a separate tool resting on the seafloor. Seismic tests happen at fixed intervals as you penetrate deeper into the earth. This means that at the end of a Seabed CPT test, you should have several depths where seismic tests occurred. This data is analysed, stacked, and migrated, ultimately giving the client a profile of the P and S wave velocities of the underlying CPT hole. Seismic CPT has a much higher resolution at shallower depths (typically less than 20m), becoming noisier as the receiver gets deeper into the ground.

Leg Penetration Analysis of Jack-Up Rigs

A jack-up rig is a temporary offshore platform that floats out to site, lowering its legs to the seafloor which then supports its weight. There are two catastrophic risks that may lead to an unplanned rapid descent of a leg after jacking up: a punch-through, and a squeeze. A punch-through is when at least one leg forces a hard geological layer through the underlying un-supporting soft layer. A squeeze is when a soft cohesive layer squeezes out between two hard layers or the leg and a hard layer. There are different variations of a punch-through and a squeeze, but all can lead to catastrophic loss time accidents. This can be mitigated, but the Rig-Move Master must know if a risk exists. This is why we conduct a “Spud Can Analysis” or a “Leg Penetration Analysis”, to determine if the ground conditions are a risk.

Figure 24: An offshore geotechnical drilling derrick (left), offshore renewable jack-up (centre), and offshore non-renewable drilling jack-up (right). Note that the drilling jack-up on the right is substantially larger than the renewable jack-up.

Spud Can Analysis or Leg Penetration Analysis is the application of both CPT and laboratory sampling data to determine if a rig leg will penetrate unexpectedly. It gives a strong profile of the geology, is typically commissioned alongside a geotechnical investigation, and can be largely be done offshore to give preliminary results. Typically, there are several ways to mitigate the damages incurred to the rig, such as jacking low to the waterline or purposefully punching through a hard layer, but this is a discussion for another article.

Figure 25: Naga 7 jack-up offshore Malaysia sinking due to a punch-through scenario or “rapid soil penetration” of one of its legs. Image courtesy of https://gcaptain.com

Who?

Living on a Ship

Living on a ship can be a very hard lifestyle. There are many things that an onshore manager can do to ensure the wellbeing of their offshore workers. Outside of fair wages, proper rest, and career progression, always be aware of food quality on a ship. Good food can make a three-month sail more bearable than a two-week sail with horrible food. Sailors constantly talk about the state of the food on a ship and within a company. Sailors will also rarely write either positive or negative “Stop Cards” for food, so it is seldom reported to onshore.

Figure 26: Food at sea can be very nutritious and well-rounded, making long stints much more bearable. This is an example of a well-rounded meal, and an excellent standard of food offshore.

Client relations

A client company will always send a representative on board the vessel to ensure that the data collected is up to standard and within the boundaries of the contract. All major decisions are made by the onshore project management teams, who are represented by the Offshore Project Manager/Party Chief (OPM/PC) and the Client Representative. The OPM/PC and the Client Rep ensure that the operation is running smoothly. They collect information from all members of the offshore team and send progress reports onshore. Generally speaking, the overall mood of the vessel is set by the relationships between the drilling team, client rep, and the OPM/PC. The best client reps/OPM/PC’s are calm, experienced, friendly, and hold realistic expectations. It is very common practice to hire freelance client representatives as most energy companies do not have dedicated employees who are experienced with the nuances of drilling operations.

Figure 27: Ekofisk 1 oil platforms in the Norwegian sector of the North Sea.

When?

Every vessel has a specific limit of wave height, and wind capacity. On most geotechnical drilling vessels, wave heights just above two meters are when work becomes treacherous in drilling operations. In the northern hemisphere, we typically schedule geotechnical work during the summer months when the weather is best. Strong winds and wave heights frequently occur above two meters in the winter, so it is best to avoid scheduling an investigation unless it is unavoidable. This difference in weather is more pronounced in higher latitudes, so the closer you are to the equator, the more working months are in the year.

Figure 28: Weather forecast for the North Sea, a commonplace for geotechnical site investigations. This forecast is for the Northern Hemisphere; in the Southern Hemisphere, the weather is poor from March to September.

Conclusion and personal opinions

A good geotechnical campaign is nuanced, yet straightforward. If proper care is taken with the sampling and testing regime, then the results can last for many years in the future. It is not uncommon to see geotechnical results from the 80’s pop up in modern geotechnical investigations. While still relevant, old investigations are less reliable, and the precedents have changed. The advent of digital equipment has made the capacity for data analysis immensely larger. Data is now manipulated by scientists to create useful frameworks which streamline interpretation times. Notable recent changes are within LPA and CPT analysis, as well as general geotechnical site investigation overview. The change to digital is not without hiccups, as generally speaking, hydraulic systems can be more reliable and easier to repair on the fly. Another general trend is to go with more CPT boreholes with fewer, more dispersed sampling boreholes. This gives the client enough engineering data to sufficiently design foundations while spending less money on laboratory testing. This is not always advised though, as offshore areas with a large variety of geology (such as buried river geologies, etc) demand higher sampling coverage.

For cable/pipeline corridors, seabed CPT’s are ideal as they can be mounted via A-frame (which comes standard on most vessels) and can provide adequate penetration. Caution must be exercised in areas consisting of thick layers of very dense sands, the seabed CPT will encounter penetration issues due to high cone tip loads. In these cases, downhole CPT may be necessary due to the driller’s ability to destructively drill through hard layers. Ideally, a cable/pipeline corridor should be investigated by a cheaper vessel equipped with a vibrocore and a Seabed CPT. While deeper investigations should be investigated by drillships.

Typically speaking, it is okay to have a higher percentage of CPT than sample locations, but it is essential to have at least one sampling location, or a generous hybrid sampling scheme. A 3/1 ratio of CPT to Sampling is a safe bet provided there is not a high variation of geology between sampling boreholes. In areas of known soil homogeneity, it is safe to have a higher ratio, even up to 6/1. During a drilling investigation, it is wise to have at least one CPT location investigated first, so those logs can be used by the driller to better adjust the density of the mud and get better sample yields.

Figure 29: the Kommandor Susan on location offshore Boston undergoing an offshore geotechnical site investigation for offshore wind farms scheduled to be built. Image used courtesy of Horizon Gardline and Hays shipping.

Acknowledgements

The author has spent the last two years conducting offshore geotechnical site investigations all over the world. This article was written with knowledge gained first-hand or through conversations with professionals directly employed in the industry. Many of the images are taken of ships and equipment owned and operated by companies within the Industry. These companies may and will hold trademarks and patent protection for this equipment. If you are an owner of this equipment and have a request for these images please notify the RASS Admin via www.rasstech.net/contact .

A thank you to the good words of Sam, Josh, Ben, Ellie, Frankie, Ben, Andy, Carl, Hesh, Patryk, Simon, Terry, Dave, Ollie, Jed, Mark, Felix, Candido, Ryan, Jason, Stevan, Jamal, Renial, Jaymar, James, Rich, Lucy, Jo, Dave, Ash, Dory, Matt, Mirna, Colin, Lance and Rich. Each of you has provided some insight and by virtue, contributed some measure to this article. A huge shout out to Horizon UK for making this all happen, and to the cooks and crew at Hays for making some damn good food.

Figure 30: The sun rising on the horizon with a wind farm in the foreground.

Image Disclaimer:

There are four types of figures in this article: (1) Images that the author took personally. (2) Images with equipment owned by companies other than RASS. (3) Images taken from the internet on well-known websites. (4) Images that have no credit or are taken from unknown origin. Images that do not have credit at the bottom of them are all (1), it is okay to copy and share these in any way. Images that are credited have been used with the express permission of the respective company and are (2). It is not permitted to replicate or use (2) images unless prior permission is taken from the respective company. Images that are taken from the internet (3) will be credited with the web address underneath, it is okay to replicate or use (3) images provided that you properly credit the respective website. If you are the owner of an image classed as (4), and would like it modified or credited, please contact RASS. It is not okay to replicate or use (4) outside the confines of this article.

IP Disclaimer:

There are four types of research contained in this article: (I) Data logs or tables that have been wiped of an origin outside of RASS. (II) Figure diagrams that were created by RASS but the methods are not owned by RASS. (III) Figure Diagrams that contain © RASS Technical. (IV) The exposition of this entire document, which was written by a non-peer-reviewed freelance scientist. No reader of this article has permission to copy or replicate for any non-educational purposes IP that would be considered (I) and (III). Sharing of this document is encouraged, and multiple .pdf versions can be produced at the