Nine Metres Under

The field notes from the first team were fine. Mostly standard material: trap coordinates, water-temperature readings, species counts, condition notes on the tagging equipment. Whoever had written them up was thorough, perhaps a little too thorough in the way people are when they are new and still think the volume of detail is the same thing as quality.

I had skimmed about forty pages in the Norwich office before I found the first warning. Page 43. A small handwritten note in the margin:

GPS receiver logging anomalous position data. Confirmed against paper chart. Discrepancy: 9.4 m beneath water surface. Assumed receiver error. Survey continued.

I read it twice.

Then I carried on. The second team's notes were in a separate binder, blue instead of green, on the same shelf. Different technicians, different dates, three weeks later in the season. I pulled it down, sat at the table with my coffee, which had gone cold, and found the same note on page 31.

Not a similar note. The same language, the same 9.4 m figure, the same stretch of channel.

One of them had even written 'assumed receiver error' in almost exactly the same phrasing. That is the sort of coincidence you get when two separate people encounter the same obvious problem and reach the same obvious conclusion. That was what I told myself, anyway.

The Environment Agency office in Norwich is a low brick building off Yarmouth Road, and on a Saturday evening it was completely empty apart from the person who had let herself in through the side door, which was me, because I had a key.

I had driven up that morning specifically to review the files before the survey. I like to know what I am walking into, and the regional coordinator had flagged the earlier GPS data in an email containing the phrase 'inconsistencies worth noting' without explaining what that meant. I hate that. Just say what it is.

What it apparently meant was that two teams had independently recorded the same equipment fault within the same eight-hundred-metre section of a remote, reed-lined channel connecting Hickling Broad with the upper River Thurne, and neither team had followed it up beyond a margin note.

I wrote 'probable calibration fault - order replacement receivers' on a sticky note and fixed it to the front of the first team's binder.

I meant it. It was the obvious answer. GPS receivers drift, especially units that have been checked in and out of a field office a dozen times without a full recalibration between uses.

Technicians notice something odd. They record it. They move on because they are on the water and have traps to pull. That is not negligence.

That is fieldwork.

The supervisor, in this case me, is supposed to catch it during review and correct it before the next team goes out. So I corrected it. I emailed the equipment coordinator from my mobile and requested three new GeoNav XR-40 survey receivers, all with current firmware and full calibration before dispatch.

Done.

I closed both binders, returned them to the shelf in the proper order, switched off the lights and locked the side door behind me.

The drive from Norwich to Potter Heigham takes under an hour if the traffic behaves. I followed the A47, joined the A149 and eventually turned on to a narrow Broads Authority access lane leading towards the slipway. I have done longer drives for worse surveys.

The road narrowed towards the end, and the vegetation pressed close on both sides. Farmland and hedges at first, then dense reed beds, willow and patches of flooded alder carr as I approached the Broad.

My vehicle was an eleven-year-old Toyota Hilux with a cracked dashboard and a passenger seat I used as a filing cabinet. It ran perfectly well. I had a petrol-station cheese and pickle sandwich, a bottle of water and the radio turned down low.

Local BBC radio faded into static somewhere along the A149 and never properly returned.

I passed a Potter Heigham village sign and found myself thinking about the area's old drainage disputes, shifting waterways and uncertain boundaries between land and water. The Norfolk-Suffolk wetland region has a long history of surveyors trying to draw permanent lines through places that refuse to stay still. I find that genuinely interesting. Almost no one else does.

Hickling Broad is the largest of the Norfolk Broads. I had surveyed it before, in different sections and different seasons. It is a good wetland.

Dense reed beds, flooded alder carr, willow roots reaching into black water and hanging lichen caught in the branches. Beautiful in a way that becomes slightly unsettling when you are alone and the light is going.

I reached the access lane shortly after nine. The village behind me had already gone quiet, which was reasonable for nine on a Saturday night, but the darkness had a particular weight to it.

There were few street lamps and fewer lit windows. I passed a fishing-tackle shop with a hand-painted 'CLOSED' sign in red. From the lane I could make out the slipway, its concrete slope disappearing into black water. A handful of houses stood back from the road with porch lights burning.

I drove past the Broads Authority access track once to look at the slipway, then pulled into a gravelled lay-by about fifty metres farther along and parked.

I had considered finding somewhere to stay in Great Yarmouth, the nearest place with plenty of late-night amenities, but it was too far out and I needed to be at the slipway by 05:00. Sleeping in the Hilux made more sense.

I had done it often enough. The back seat was not comfortable, but it was nearly flat if I folded the front seats forward and laid out a sleeping bag, which I did. I set an alarm for 04:30, giving myself roughly seven hours. I was confident that would be enough.

The survey was planned as a single-day operation on 15 November 2026. Nine monitoring stations were scheduled. We would pull the existing traps left by the previous teams, record any new captures, retag and release them, and document the GPS data properly. With the replacement receivers, we should be back at the slipway by three or four in the afternoon. I would have clean data entered by 17:00.

Then I would return to Norwich with a closed file and one less problem on my desk.

That is the thing about a problem that looks simple. You stop looking at it.

I lay in the back of the Hilux with the sleeping bag pulled up, my mobile on the seat above my head and the windows cracked by a couple of centimetres. The air was damp and close, the way it becomes beside the Broad.

Frogs called somewhere beyond the reeds. Water moved in the distance. It was the kind of quiet that stopped being quiet as soon as you listened to it.

I thought about the margin notes again for perhaps two minutes.

Two teams. Same coordinates. Same error. Same figure: 9.4 m beneath the water surface.

An old drainage-board and waterways survey chart held in the Environment Agency archive showed an 11.3 m hollow at the deepest surveyed point in that section of channel. That did not make either team's reading sensible. A GPS receiver placing its user 9.4 m beneath the water was not measuring water depth. It was malfunctioning.

Receivers do that. They lose satellite lock. They interpolate badly. They produce figures that mean nothing.

You replace the receivers and the problem goes away.

I had replaced the receivers. The problem would go away.

I closed my eyes.

The frogs continued. Somewhere on the Broad, something moved in the water. One splash, then nothing.

It could have been anything. Hickling has pike, eels, otters and even large terrapins abandoned years ago and grown patient in the mud. I knew the Broad. I knew what lived in it.

I fell asleep certain I would be home by evening.

My mobile alarm sounded at 04:30, but I was already awake. I had slept for perhaps three hours. The Hilux was warm and slightly stuffy, and I had been lying there listening to the frogs for some time before the alarm began, staring at the roof lining.

I got up and worked through my gear checklist on the tailgate by head torch.

Two waterproof notebooks. Flagging tape. Measuring equipment. Three new GeoNav XR-40 survey receivers: one for me, one for the university technician and one backup clipped to the outside of my buoyancy aid, where I could see the display without stopping.

All three were logging. All three showed the same correct coordinates for the Broads Authority access track off the A149, exactly where I was standing. Their elevation readings were normal.

I checked them twice because that was the protocol, and because I had written 'probable calibration fault' in a margin the previous evening and wanted to feel confident about the replacements.

They looked fine.

I put on my buoyancy aid and drove down to the slipway. It was still dark. There were no lights except mine and a faint grey edge beginning beyond the reeds to the east. It was not daylight yet, merely a sky becoming slightly less black.

The water was flat and reflected nothing. It seemed to absorb the headlights where the concrete met the Broad. I switched off the engine and sat for a moment.

Another vehicle was already there, an older Toyota Hilux with a University of East Anglia parking permit in the rear window. That was right. Elliot Price had been contracted through UEA.

He was already at the water. He had the aluminium fishing punt half loaded and was moving equipment from the back of his Hilux with the calm efficiency of someone who had done it before. I watched him for a moment before getting out.

The cool boxes were positioned properly, their weight low and centred, which not everyone does without being told.

I climbed out and introduced myself.

'Elliot, yes. Dr Amina Malik mentioned you.'

He shook my hand, firmly and briefly, then returned to the cool box he was moving. I liked that. I have worked with technicians who want to talk through the entire morning before we are even on the water, and I am not built for that at 05:00.

He wore a yellow high-visibility buoyancy aid over a grey field shirt and already had a head torch on. It had a decent beam, not one of the cheap ones that fails after two hours.

His gear was organised sensibly on the tailgate. Data notebooks together, sampling equipment in one container, personal kit in another. He had clearly read the preparation sheet I had emailed, or at least most of it.

I checked the equipment manifest against what he had loaded while he finished with the cool boxes. Everything matched. I initialled the form and tucked it into my buoyancy aid.

About ten minutes later, headlights appeared along the access lane.

A Broads Authority vehicle came down towards the slipway. Gareth Morgan had been assigned to us under the joint survey permit, standard practice where field operations crossed the responsibilities of the Environment Agency fisheries and ecology team and the Broads Authority.

You get a ranger for the safety inspection and to cover the jurisdictional overlap. I had worked with Broads Authority rangers before. They were generally fine. Some were very talkative.

Gareth Morgan was very talkative.

He was a large man, broad through the shoulders, with closely cut salt-and-pepper hair. He moved easily for his size and never seemed hurried.

He shook my hand and said, 'Harriet Shaw, Environment Agency. I know your name. I read the permit.'

Then he shook Elliot's hand and said, 'You're the university lad.'

Elliot said, 'Yes, sir.'

Gareth appeared to find this acceptable.

We conducted the safety inspection together. Gareth had a clipboard and I had mine, while Elliot stood at the bow and answered questions as we asked them.

Buoyancy aids, checked. Distress-flare kit, checked. First-aid kit, radio and emergency-contact procedure, laminated and cable-tied beside the controls, all checked. A small Broads Authority rescue launch, carried in tow for emergencies, was secured behind the survey punt.

Gareth worked through the list without rushing. I appreciated that. Some people treat a safety inspection as a formality. Others treat it as if it matters.

He treated it as if it mattered.

While we worked, he talked. That was simply how he operated. Talking and working at the same time, seamlessly, as though he had been doing both for decades.

His Norfolk accent was slow and comfortable. He talked about the Broad, the reed-lined channels and the water level being slightly lower than the previous year. Then he mentioned Silas Fenwick.

'You know Silas?' he asked while checking the distress-flare kit, without looking up.

I said I did not.

'Old fishing guide. Been taking people out here thirty years, perhaps longer. Knows Hickling better than anyone. Good man.'

He snapped the kit closed.

'Won't take clients into that upper channel after October. Not for any fee. I've seen people offer him twice or three times what he normally charges. Doesn't matter.'

Elliot looked up from the bow.

'Why not?'

Gareth shrugged, but it was not a dismissive gesture. It looked more like the movement of a man reporting an established fact.

'Says it gets occupied.'

I wrote something in my notebook. I cannot remember what. Something routine.

It gets occupied.

That was the phrase. I heard it, filed it somewhere at the back of my mind under local colour, and carried on working.

I had been in field ecology long enough to know that every region has its version of this. The stretch of river you do not navigate after dark. The marsh where you never wait alone. The old guide who tells you there is something wrong with a particular place if you buy him a cup of tea.

It is not necessarily meaningless. Sometimes it begins with a real observation: unusual currents, submerged structures or patterns in animal behaviour that become a story through repetition. But it is also the sort of warning people repeat until it sounds like more than it is.

Silas Fenwick had worked the Broad for thirty years. That was plenty of time to develop opinions. It was also plenty of time to build a reputation that kept his diary from becoming too full during the quiet months.

I did not ask what 'occupied' meant. I had nine monitoring stations to complete, with water-temperature and dissolved-oxygen readings to record at each one. I had GPS coordinates to verify against the earlier teams' data, and I wanted to be back at the slipway before the rapidly failing late-autumn light became darkness.

I had a full day.

Gareth completed the safety inspection and signed the bottom of his copy. I signed mine. Elliot signed where I indicated.

We launched shortly before sunrise.

The sky was becoming a flat grey-orange at the horizon. There was no warmth in it yet, only colour. The water was black. Genuinely black.

Not dark blue or dark green. Black.

Hickling can look like that in late autumn, when tannins darken the water and the light comes in low. The aluminium punt moved away from the slipway, and the water closed behind us almost without sound.

No returning ripple. No reflection.

The alder carr and reed beds along both banks were only shapes, dark against a slightly paler darkness. Gareth controlled the outboard from the stern. Elliot sat at the bow with his head torch. I was amidships with my clipboard and one GeoNav receiver clipped to my buoyancy aid.

It was already running and logging. I checked the display. Correct coordinates, correct elevation, everything exactly as it should have been.

The channel opened ahead and the reeds pressed close on both sides. The outboard settled into a low, steady drone. Somewhere among the alders, a bird called once and stopped.

The air smelt of mud, water and something faintly vegetal, the smell of standing water that has been undisturbed for a long time.

I watched the GPS display and thought about the clean data the replacements would provide. I imagined writing 'resolved - equipment fault' in the follow-up report and closing the earlier anomalies for good.

The channel drew us in, and the slipway vanished behind us.

The notebook was a standard waterproof field book, the same sort we used for every survey. Yellow cover, wire-bound, water-resistant paper. I had handed it to Elliot at the first monitoring station at 06:40 and had not thought about it since.

By mid-morning we had cleared eight stations and were preparing the ninth. I asked him to read back the last four sets of coordinates so I could cross-check them against my receiver log.

He handed over the notebook without looking at me. I assumed he was tired.

It was after ten. The sun was up, although it offered little warmth, and the channel had narrowed enough that Gareth had to cut the outboard and use a pole to work us through willow roots and submerged alder branches.

I turned back through the last few pages and ran my finger down the columns.

The first four stations were fine. Neat handwriting, the figures upright, tight and evenly spaced. The column headings were underlined twice, exactly as I had seen on Elliot's contractor paperwork in the Norwich office two weeks earlier.

He had completed those forms in ballpoint pen at the corner of my desk, and I had watched him without really watching, the way you half-notice things while concentrating on something else. But I had seen enough to know his hand.

It slanted to the right. The letters were rounded at the tops, with consistent pressure.

Stations five to eight were different.

The writing was cramped and slanted to the left. The letters were squeezed together and tilted the wrong way, as though someone had been told what letters looked like but had never practised holding a pen.

The GPS columns for stations five, six, seven and eight were blank. They had not been crossed out or marked 'error'. There was no warning or explanatory note. There were simply empty boxes where decimal coordinates should have been recorded to four places.

I stared at the page for some time.

Sunlight came through gaps in the alder carr in flat strips and struck the water at an angle that erased everything beneath the surface. I could not see the bottom. I could not see much of anything except the notebook in my hands and the two men in the punt with me.

Gareth was at the stern, working the pole. Elliot sat at the bow with his back to me, watching the channel ahead.

I told myself his receiver might have failed. Receivers fail. Batteries die. Firmware crashes. A unit is knocked against the gunwale and damaged. I had seen receivers fail in the field, leaving a technician with no choice but to continue.

That happens.

But when it happens, you write something in the column. You write 'receiver failure' or 'no signal', or you put in a dash and circle it. You do not leave the boxes empty and change your handwriting.

I closed the notebook and set it on my knee.

I looked at the back of Elliot's head: sandy blond hair, yellow buoyancy aid, one hand loosely holding the bow rope. Everything about him was normal. There was nothing I could point to.

I opened the notebook again and compared stations five to eight with stations one to four. Then I studied the empty GPS columns.

Contractor paperwork is boring, and people fill it out the same way every time. Handwriting is consistent because it is a habit. You do not decide how to form a lower-case 'a' whenever you write one. Your hand simply does it.

Elliot's lower-case 'a' in the first four entries had a small, closed loop at the top. The 'a' in the sixth entry was open and almost printed, as if someone were copying a letter it had seen but did not properly understand.

I am not a handwriting analyst. I want that to be clear. I am an Environment Agency field ecologist who surveys wetland fauna across the Norfolk-Suffolk wetland region. I know a great deal about tagging protocols and GPS calibration and almost nothing about forensic document examination.

But I knew what Elliot Price's handwriting looked like.

This was not it.

I needed one thing I could establish with certainty before I did anything else.

'Just remind me who your supervising professor is at UEA,' I said casually, keeping my eyes on the notebook. 'I need to update the contact section in my report.'

He turned round.

That was normal. It was what anyone did when spoken to.

'Professor Wycliffe,' he said. 'Peter Wycliffe.'

I wrote the name in the margin of my own notebook and nodded as though I were recording it.

'Got it. Thanks.'

He turned back towards the bow.

Professor Peter Wycliffe had retired from UEA in 2020. I knew because I had rung the departmental administrator while assembling the contractor file. She had told me Wycliffe was no longer there and suggested two current members of staff who could provide contractor references. I had written both names in the project file.

I also knew, because the administrator had mentioned it during the same call, that Professor Wycliffe had died in 2022.

The thing sitting at the bow of my punt had given me the name of a dead man and then smiled.

It did not smile because it was nervous. It did not smile because it realised the answer was wrong and wanted to cover the mistake. It smiled as if it had completed a task.

Clean and empty. No recognition behind it at all.

I said nothing.

I looked at Gareth, roughly four metres behind me at the outboard. He was watching the channel and pushing the pole into the silt to move us forwards.

He had not heard. Water and reeds deadened sound, and a short conversation at the bow did not carry well to the stern.

I placed my personal mobile in my lap beneath the notebook. Starting from the first page, I photographed every relevant page.

Twelve images.

I kept the movements small, using only my thumb on the screen, holding the notebook at an angle as though I were reading it. The thing at the bow did not turn. Gareth was concentrating on steering us around a submerged branch on the port side.

No one was watching me.

I captured all twelve pages. Then I opened the voice-recorder application, began a recording and placed the mobile face down on my knee.

The technician was humming.

I had noticed it a few minutes earlier and dismissed it as background noise. People hum when they are bored or when they are not thinking about anything. But now that I listened, it was not casual.

The same four or five tones repeated. It was not a song or anything I recognised, only a pattern looping back on itself at exact intervals.

I timed the repetitions.

Four seconds. Then four seconds. Then four seconds again.

Every time.

People do not hum like that. Even if you repeat the same phrase, you drift. You speed up or slow down. You miss a note. The intervals are not mechanical.

This was mechanical.

I let the mobile record for thirty seconds, then stopped it.

I needed to warn Gareth without saying anything in front of the thing at the bow.

I opened my messages and entered his number from memory. I had had it for three days, since the survey briefing, and I am good with numbers.

Stay alert, I wrote. Don't ask why yet.

I watched him from the corner of my eye. His mobile buzzed in his breast pocket. He took it out, read the message and looked at me.

I lowered my eyes to the notebook.

He put the mobile away and continued poling.

I had nothing else to give him. I did not have a word for what I was looking at.

I had spent two weeks certain that the earlier teams' GPS failures were a calibration problem. I had ordered replacement receivers. I had driven to Potter Heigham on a Saturday night and slept in my Hilux because I believed this was a solvable equipment fault and wanted to arrive early to confirm it.

The real Elliot Price had driven to the slipway that morning. His Hilux proved that. He had arrived before me and before Gareth because he was the kind of technician who came early and began loading equipment without being asked. His references had said as much.

But I had not seen him walk from his Hilux to the slipway. By the time I arrived, someone wearing his face was already completing the loading.

If the real Elliot had disappeared that morning, it had happened somewhere between his vehicle and the water.

I sat with the notebook on my knee, my mobile in my pocket and the empty GPS columns in front of me. I thought about the black water at the slipway and the way it had closed behind us without sound.

I thought about all the isolated cuts, dykes, reed beds and flooded carr between Norwich and Hickling Broad, places where a vehicle or a body might remain unseen for days if no one knew to look.

The technician at the bow continued humming.

Four seconds.

Four seconds.

Four seconds.

I opened my field notebook to a clean page and wrote the date, time and coordinates displayed on my receiver. I wrote carefully in my own hand, right-slanting and rounded at the tops.

Then I looked at the blank columns in Elliot's notebook and copied the coordinates for all four missing stations into mine, because the data still had to exist somewhere.

I continued working.

The channel moved beneath us. Reeds and flooded alder carr pressed close on both sides.

The outboard stopped at 14:47.

I know the exact time because I had been recording timestamps on everything since mid-morning. It was habit, but it was also because some part of me knew I was constructing a record.

I did not allow myself to think beyond that.

One moment the engine was running, producing the low, reliable thrum of a four-stroke outboard that you cease hearing after the first hour. The next moment it was silent.

Gareth muttered something and pulled the starter cord twice. Nothing. He checked the fuel line and kill switch, then tried again.

'It's not flooded,' he said. 'Line's clear.'

I was already looking at my GeoNav XR-40.

It was the new receiver I had requisitioned after writing 'probable calibration fault' in the margin of the earlier team's records. Brand new, fully calibrated, current firmware. I had charged it in the Hilux the previous evening.

It showed our current position as 9.4 m beneath the surface of Hickling Broad.

I stared at the display.

The screen was undamaged. The receiver was functioning. The satellite count in the upper corner showed nine, which was a strong lock.

Nine satellites, and the receiver said I was sitting 9.4 m beneath the water in an aluminium punt in the middle of the afternoon.

I took out the old drainage-board and waterways survey chart I had copied from the Environment Agency archive. It came from the 1957 drainage-board survey and remained the most detailed depth record for that section of the channel.

I found our position.

An 11.3 m hollow was marked with a small asterisk and a note I had read three times without attaching any significance to it:

Anomalous substrate return. Repeat survey recommended.

No repeat survey had been recorded. The recommendation had remained unresolved since 1957.

We were directly above the hollow.

I did not tell Gareth immediately. I was still trying to work through it.

Both earlier teams had recorded GPS anomalies within this same eight-hundred-metre reach. The same coordinates, different teams, different equipment and different dates. In Norwich, I had called it calibration error. User error. A routine equipment problem that could be corrected without investigation.

Order new receivers and move on.

I had written it in my own hand: 'probable calibration fault'.

Their equipment had not been faulty.

They had been here. Exactly here. They had seen the same 9.4 m negative elevation I was seeing, and I had answered their records with a requisition form and a reminder about calibration procedure.

That knowledge settled in my chest in a way I did not want to examine, so I did not.

I returned the chart to my buoyancy aid and looked up.

The technician was standing.

He was not half-rising to shift his weight or look at something. He was fully upright in an unanchored aluminium fishing punt rocking gently in the channel, which was not something a trained survey technician did.

You do not stand unnecessarily in a small flat-bottomed punt on open water. It is one of the first rules you are taught. Elliot had completed the required field-safety programme. I had seen the certificate in his contractor file.

He stood at the bow with his hands by his sides, looking down at the water.

'Hey,' I said.

'Sit down.'

He did not sit.

He stepped on to the gunwale, rocking the punt hard enough that I seized the starboard rail with both hands. Gareth shouted from the stern.

The technician balanced on the gunwale for perhaps a second.

Then he stepped off on to the water.

Not into it.

On to it.

He stood there with both feet on the surface of the channel, his weight fully settled, the water supporting him as if it were a jetty.

I watched it happen from less than two metres away. I have no means of turning it into anything else.

He stood on the water for approximately two seconds. I counted because that is what I do.

I count.

Then the water accepted him.

That is the only word I have for it. Accepted, as if a decision had been made.

He went straight down. He did not fall or flail. He simply descended, and the channel closed over his head.

No bubbles rose.

I waited. You always get bubbles: air trapped in clothing, air expelled from the lungs, something.

I watched the surface for fifteen seconds and saw nothing. No widening ring where he had entered. No disturbance at all.

The water lay flat and black, as though the preceding two seconds had not happened.

Gareth was already moving towards the bow.

'Man overboard,' he said, because that was his training.

He leant over the side, peering into water that showed him nothing.

I looked at my GPS receiver.

It still placed me 9.4 m beneath the surface.

I thought about the previous teams. I thought about my note in the margin. I thought about Elliot Price, who had driven to the slipway before sunrise, and the pages filled in with cramped, left-slanting handwriting that was not his.

I understood that whatever had occupied the bow of the punt since we launched had not been Elliot Price.

Perhaps the substitution had happened at the slipway, somewhere between his Hilux and the water. Perhaps it had happened earlier. I did not know.

I did not know where the real Elliot Price was.

I still do not.

I put the receiver into my buoyancy aid and picked up a pole.

'We need to move,' I said.

Gareth looked at me. He was still bent over the bow.

'We can't just-'

'There's nothing down there for us to find,' I said. 'We need to move.'

He did not argue. I think he had seen the surface too.

We poled the disabled survey punt east for roughly forty metres, working against a light current I had not noticed while the outboard was running. Flooded alder carr pressed close along the northern bank. We nosed the punt among the roots and secured it to a thick alder root.

Hanging lichen and willow branches enclosed us on both sides. The air smelt of mud and something older than mud.

Gareth tried the Broads Authority control-room frequency first and received static. He switched to the emergency channel and caught something that might have been a voice before it dissolved into noise. The signal had been poor all day. Now it was worse.

'I'll keep trying,' he said.

I nodded.

I was looking up the channel.

A figure stood in the water about sixty metres north of us, beside a derelict timber shooting hide I had marked on the chart earlier. The structure was half-collapsed, its timber grey and soft-looking beneath old black creosote, with dry reed litter banked against one side.

From the surface and the surrounding reed stems, the water beside the hide appeared to be approximately 1.2 m deep. The figure stood as a person might stand in 1.2 m of water: upright, comfortable, apparently resting on the bottom and untroubled by the current.

But the chart placed the 11.3 m hollow beneath that reach, and my GPS receiver still indicated a position 9.4 m beneath the water surface.

The figure wore a yellow high-visibility buoyancy aid.

I watched it for a long time without moving.

The late-autumn light was failing quickly. The sun had dropped behind the reed line, flattening every colour and softening the edges of things.

The figure did not move.

It stood there as though it were waiting, or as if it had already been waiting for a long time and was in no hurry.

I thought of a large terrapin resting on the bottom of dark water, motionless for hours, so patient that divers might mistake it for a stone or a piece of timber until it moved.

I did not look away from the figure.

'Gareth,' I said, keeping my voice level. 'Don't look up the channel yet. Keep working on the radio.'

He did not ask why. He returned to the handset.

I kept watching the yellow buoyancy aid and the figure that had sat at the bow all day, humming in exact four-second loops, writing with the wrong hand and naming a dead professor when I asked a simple question.

It had stepped off the punt, stood on the surface and sunk without bubbles. Now it was sixty metres away, upright beside the shooting hide where the water looked 1.2 m deep although the receiver still reported 9.4 m beneath the surface.

Patient as anything, wearing the buoyancy aid I had watched Elliot Price wear at the Broads Authority slipway at 05:00.

I took my mobile from my pocket and photographed the figure, then took another photograph. The light was poor, and I could not tell whether either image had captured it clearly.

I kept my eyes on the channel.

I could not find the words quickly enough. That is what I keep returning to.

I had my mobile. I had twelve photographs of notebook pages. I had a thirty-second audio recording of the humming and my own written coordinates.

But when Gareth finally looked up the channel and said, 'Is that your technician?' I had no sentence ready.

I had spent the afternoon building a case in my head, cataloguing evidence and separating what I knew from what I could prove. None of it prepared me to say aloud what I thought was standing sixty metres away, in water that appeared shallow but should not have been, wearing Elliot Price's buoyancy aid.

So I said, 'Gareth, wait.'

He was already preparing the small Broads Authority rescue launch.

He was not being reckless. That is the part that is difficult to live with.

He was doing exactly what a Broads Authority ranger was supposed to do after seeing a person apparently stranded in open water while the light deteriorated. He moved quickly and calmly, the way someone moves when he has carried out water rescues before and understands that the useful window is short.

The disabled survey punt remained secured to the alder root. Gareth climbed into the separate rescue launch we had towed for safety, released its line and started the small outboard before I had managed to move off the bank.

I called his name once.

He looked back with an expression that said, 'I've got it.'

Then he turned the launch up the channel.

I went after him on foot along the bank and through the shallows. The mud beneath my boots was soft and uneven, full of willow roots and flooded alder tangles, and I had to watch every step or fall.

The water was cold. Not yet cold enough to cause hypothermia, but cold enough for me to feel it through my trousers the moment it rose above my ankles. I kept Gareth's rescue launch in the edge of my vision and the yellow high-visibility buoyancy aid directly ahead of me, moving as quickly as I could without losing my footing.

We had secured the disabled survey punt to an alder root before Gareth took the small Broads Authority rescue launch that we carried for emergencies. Now the channel curved slightly to the left beyond a stand of flooded alder carr, and the derelict shooting hide came fully into view.

It was badly rotten. Half its slats were missing. The platform leaned at perhaps thirty degrees, and the entire structure rested on four posts that had stood in the water so long they had turned black. It was the sort of place that had been abandoned for so many years that ownership no longer seemed to apply.

The figure was still standing beside it.

Gareth cut the outboard about fifteen metres away and allowed the rescue launch to drift. I was perhaps forty metres behind him, still forcing my way through the apparent shallows. I had my torch out.

I do not remember unclipping it from my buoyancy aid, but it was in my left hand. I was using the light on my mobile in my right hand to watch my footing, trying to close the distance without straying into water deep enough to slow me further. Gareth's launch drifted to within eight metres of the figure.

He called something.

I could not make out the words, only his voice and its professional, level tone. The figure turned its head towards him. It was the first time I had seen it move since it had sunk into the channel without a bubble.

I reached a point about twenty metres away and stopped. The bank fell sharply there, and within two steps the water would have risen from my knees to my chest.

I switched on the torch and aimed it along the channel.

The technician's face was turned towards Gareth. The yellow buoyancy aid was unmistakable, as was the sandy blond hair, now dark with water. It was the face I had watched across the aluminium survey punt all day. The face that had smiled at me after casually naming a professor who had retired in 2020 and died in 2022.

It was standing above water that looked no more than 1.2 m deep, although the old chart showed an 11.3 m hollow there and my GPS receiver still placed the position 9.4 m beneath the surface.

Both its hands were inside Gareth Morgan's chest.

They had not entered through a wound or by any means that made structural sense. Both hands were buried beyond the wrists, moving with calm, methodical precision.

It was like watching someone search a bag for an object they knew was there. Patient. Unhurried.

Gareth was still upright. He was making sounds, but not words and not screams. His arms hung at his sides, and he was breathing in quick, shallow gasps. I could see his shoulders moving from twenty metres away.

I did not move for a second. Perhaps two.

Then I raised the torch and directed the beam straight into the technician's face.

The light did not come back.

It went in. When the beam struck its eyes, it visibly dimmed, exactly as a torch does when its batteries are failing. My batteries were not failing. The dimming happened only when the beam crossed its eyes.

It was as if the light were being drawn inside and nothing were emerging from the other side.

I have thought about that many times since. I am a field ecologist. I spend a great deal of time thinking about eyes: the tapetum lucidum, reflective layers, eyeshine in nocturnal species. I know what eyes do with light. I know what they do not do.

What those eyes did to my torch beam is not something eyes can do.

Standing there with the water at my knees and the light fading into its face, I understood something I should have understood earlier.

The thing had not arrived when we did. It had not followed our punt along the channel. It had been lying completely still in the 11.3 m hollow when we passed over it that morning.

It had waited like a large terrapin on the bottom of dark water, an animal capable of remaining motionless for hours until someone reached down and touched what they had mistaken for a log.

It had been there before either of the earlier teams. It had been there when both surveys recorded the same impossible GPS elevation at the same coordinates in the same remote, reed-lined channel connecting Hickling Broad with the upper River Thurne. It had been there when I wrote 'probable calibration fault' in the margin and ordered new receivers.

It had been patient in a way that had nothing to do with hunger or urgency. It had simply waited for whatever condition it required.

And Elliot Price, whom I had contracted through the University of East Anglia, had driven his eleven-year-old Toyota Hilux down the Broads Authority access lane before sunrise on 14 November 2026. He had arrived before me. He had left the Hilux and started towards the slipway.

But he had never reached the water.

Whatever had been in that hollow had met him somewhere between his Hilux and the slipway. Whatever happened there, the thing had then finished loading our equipment in the dark and joined us wearing his form.

I do not know where the real Elliot Price is. I did not know then, and I do not know now. That is the part I cannot make fit into any formal report.

Gareth's rescue launch overturned.

I do not know precisely when it happened. I think that while I stood there holding the torch on the figure, the launch drifted close enough for Gareth's shifting weight to roll it. I heard the aluminium hull slap against the water.

When I looked, the launch was upside down and drifting slowly back along the channel towards me.

I waded towards it. The water reached my chest and then, for three steps, my neck. I caught hold of the hull and pulled myself along its side. A waterproof orange bag remained clipped to a D-ring inside the overturned launch, floating beneath it.

I knew what it contained before I released the clip. Gareth carried a single-shot marine distress-flare pistol in his approved safety kit. I had seen it during the pre-launch checks that morning.

I opened the bag. The pistol was there, with one flare already loaded. Gareth had loaded it at the slipway.

It was the only weapon I had.

I stood with the water against my chest, the flare pistol in one hand and the torch in the other.

The figure had not moved. Gareth had fallen silent.

I did not know whether a flare would harm the thing. The hide, however, was made from timber that had been rotting and drying for years. Its boards were coated in old creosote, and dry reed litter had collected across the tilted platform. It stood less than three metres behind the figure. It was the only target available to me.

I waded towards it.

The bottom was uneven. I moved slowly with the flare pistol raised while the water pressed against my chest. When I was less than three metres from the rotten slats, I stopped and planted my feet as firmly as I could in the mud.

I fired directly into the timber.

The hide caught at once.

That is the only way I can describe it. The flare struck the wood, followed by a fraction of a second in which nothing happened and I thought it had failed. Then the entire left side ignited.

Dry rot burns differently from green wood. There was no warning cloud of smoke. It simply went.

The thing wearing Elliot's buoyancy aid turned towards the fire. It did not start or flinch. It turned as a person might turn after hearing their name spoken from across a room: slowly and deliberately, as if the flames were only mildly interesting.

In that same moment, Gareth Morgan went under.

I did not see it happen. I heard it.

It was not a splash. It sounded more like the water accepting a weight, a heavy and strangely soft sound. When I swung the torch towards him, Gareth was gone.

The rescue launch was still there, overturned and drifting in the heat pouring from the hide. Gareth was nowhere on the surface. He had simply vanished.

I dropped the spent flare pistol.

I do not know where it landed. I went under.

The water was completely black. I still held the torch, but it did almost nothing. Its beam ended less than a metre from my hand, as though the water itself absorbed the light.

I kicked downwards and swept the beam in a broad arc. There was no current dragging anything away, no body and no shape in the darkness. My fingertips touched soft mud at the bottom, and I pushed myself back to the surface.

I drew one breath and dived again.

The second dive was the same, although I may have gone deeper. Without a point of reference, it was impossible to judge. I moved the light across mud, tangled willow roots and a rusted object I could not identify. It might have been an old anchor or a piece of drainage equipment left there decades ago.

There was no Gareth Morgan. There was nothing I could close my hands around.

When I surfaced for the second time, the burning hide was collapsing into the channel. Its right side gave way all at once. The slats struck the water, and fire spread briefly across the surface, feeding on old creosote and whatever else had leaked from the timber. Then it began to die, section by section.

The smoke was thick. It smelt of creosote and something else for which I had no name, something that made the back of my throat contract.

I kept the torch moving across the surface, searching for Gareth and for the thing.

Neither was there.

The channel was only a channel again: black water, flooded alder carr at the margins, willow roots and dense reeds, hanging lichen, and the noise of wreckage settling as it burnt. The figure in Elliot's yellow buoyancy aid had disappeared.

I thought it might have returned to the hollow, but I had no proof. It could have gone elsewhere. I did not know whether it had an elsewhere.

Gareth had gone with it.

I remained there with the water against my chest for perhaps a minute, perhaps longer. I am not a reliable judge of time when my mind is doing what it did then. I kept replaying the sequence: the flare striking the wood, the figure turning, Gareth going under. I tried to find an alternative order in which Gareth might still have been on the surface.

There was none.

I had made the only choice available to me. The fire had pulled the thing's attention from Gareth for a moment, and the water had taken him. The flare had worked as well as it possibly could, and it had not been enough.

That took time to settle inside me. I will not attempt to describe what it felt like, because I do not think I have the words and I am not certain the feeling is the point.

The point is that I was alone in the channel after dark. The spent flare pistol lay somewhere beneath the surface. Gareth Morgan was dead or missing in a place from which no ordinary person could have vanished, and the thing that had worn Elliot Price's face throughout the survey was no longer visible.

I was the only member of the crew left.

I pulled my field notebook from the waterproof case in my buoyancy aid. It was standard Environment Agency equipment, issued for precisely this kind of emergency. By the light of my torch, I found the laminated emergency protocol inside the front cover.

I had read it during my induction review at the Norwich office. I knew what it said, but I read it again.

'Lead ecologist must not leave the site until the police have arrived and received a full briefing. Departure before police contact constitutes a breach of field safety protocol and may result in disciplinary review.'

I read it twice while smoke from the collapsed hide drifted over the water.

The protocol made sense. It still makes sense. When something goes wrong in the field, the lead ecologist must remain to explain events, guide police officers through the scene and protect the chain of information. Leaving means reconstructing everything later from memory, somewhere else.

I had written a version of that protocol myself two years before for another district. I believed in it.

I returned the notebook to its case. Then I looked at the channel and the late-autumn darkness closing around it.

True darkness comes rapidly in November, especially beneath flooded alder and dense reeds after the sun has dropped. The GeoNav XR-40 in my buoyancy aid continued to report the same coordinates and an elevation 9.4 m beneath the water surface. The radio from Gareth's launch was underwater with the overturned vessel.

My mobile found a signal perhaps thirty per cent of the time in that part of the Broads. At that moment it had none. I did not know precisely where the nearest usable slipway was. It was a long way by water and might have been shorter over land, if I could find a route through the reed beds and carr.

The thing was probably still nearby.

I did not know whether it remained within the channel or could move beyond it. I did not know what it did when it was not standing over an impossible depth in a dead man's buoyancy aid. There were many things I did not know.

I knew that Gareth Morgan had taken a rescue launch towards it and disappeared.

I swam for the nearest bank.

It took longer than it should have. Until I began swimming with purpose, I had not understood how cold the water had become. My arms slowed.

Hypothermia does not announce itself. It merely makes every action require more effort.

I kept my strokes even, refused to rush and eventually reached the bank. I gripped a willow root and pulled myself onto a narrow shelf of mud and dead vegetation. For a minute I lay there and breathed.

The mud smelt of sulphur, rotting leaves and something organic I did not want to identify.

Then I stood and began moving along the bank.

The aluminium fishing punt lay about three metres offshore, tied to an alder root with nylon rope. It was perhaps four metres long, with a small outboard tilted clear of the water. Its transom had been painted red, and an obvious metal patch covered the port gunwale.

There was no equipment aboard that I could see. It was simply tied there, waiting.

I stopped when I recognised it.

It was Silas Fenwick's punt.

Gareth had described it that morning at the slipway: the red-painted transom, the patched port gunwale, the battered aluminium hull Silas had used for years. Silas had spent three decades working on Hickling Broad. He was the guide who called the phenomenon 'getting occupied' and would not take clients into that remote channel after the November restriction began.

Silas had known to leave the punt there.

That was what stayed with me as I looked at it. He had not guessed or hoped. He had known someone might need a boat on that bank, on that side of the channel, after dark in November. He had known it well enough to tie off the punt and walk away.

He had also known not to remain aboard it.

I untied the rope and climbed in.

After that, my memory is fragmented.

I remember the aluminium hull scraping against something in the dark. I remember trying to lower the outboard. I remember my hands shaking so badly that I could not tell whether I had gripped the starter cord. Exhaustion and hypothermia had reduced the night to disconnected moments, and I cannot say how far I travelled or how the punt reached the opposite bank.

Then I remember mud.

Cold mud, smelling of decay and standing water and something beneath both for which I have no word. My hands were buried in it to the wrists. My face was close to the surface.

I was on the Potter Heigham side of the channel, and the sky was turning grey at its edges. It was the flat grey before dawn, too weak to provide useful light but enough to promise that light was coming.

I did not lose consciousness in a way I can clearly identify. I simply stopped moving. My body made the decision without consulting me.

I collapsed into the mud at the bank and remained there. I do not know for how long.

At some point a pair of heavy boots stopped about half a metre from my face. I could see the stitching around one toe.

'Hello,' a voice said. 'Can you hear me?'

I answered, although I do not know what I said. Probably my name. That is what people do when asked if they can hear: they prove it by identifying themselves.

He crouched beside me. He had a square jaw, auburn hair and green eyes. His identification marked him as Daniel Hales, a Norfolk Constabulary rural and waterways officer. One hand rested on my shoulder, and he looked at me in the way experienced officers look at a serious collision: steady and professional, already working through the checklist in his mind.

'I'm Daniel Hales,' he said. 'I've got you. Don't try to move yet.'

I had no intention of moving. It sounded like too much work.

He called the control room on his radio. I caught fragments: adult female, hypothermic, possible trauma, Broads Authority access track off the A149.

He gave them the coordinates, requested the East of England Ambulance Service and asked for an urgent response. Then he crouched beside me again and asked for my name.

'Harriet Shaw. Environment Agency fisheries and ecology team.'

He nodded as if that explained something. Perhaps it did.

'Are you injured?'

I considered the question. My hands were cut. So were my arms. Something along my left side had been hurting for some time, although I had been ignoring it.

'I'm not sure.'

He carried out a quick assessment, checking for bleeding and anything immediately life-threatening. He found the lacerations on my forearms and another on my neck that I had not known was there.

'These look like barbed-wire cuts,' he said.

'There wasn't any barbed wire.'

'All right.'

He did not press me. Daniel Hales was very good at not pressing.

The ambulance arrived about twenty-five minutes later. I know because I watched the sky grow steadily lighter while we waited.

Daniel remained beside me and kept me talking. He asked practical questions. How many people had been aboard the punt? When had we launched? Were my emergency contacts stored in my mobile?

He was keeping me conscious and gathering information at the same time, without making either task feel urgent.

I told him what I could.

There had been two other crew members: Elliot Price, a UEA technician working as a contractor, and Gareth Morgan, a Broads Authority ranger assigned under the joint survey permit. Both were missing.

Daniel wrote their names down. He did not ask me to explain. Not then.

James Paget University Hospital in Gorleston-on-Sea was roughly forty minutes away. I drifted in and out during most of the journey.

The East of England Ambulance Service paramedic treating me looked about twenty-three. He kept talking about nothing in particular: the weather, the road and whether I had been to Great Yarmouth before. I think he was filling the silence so that I would not disappear into it.

Mostly, it worked.

The hospital was warm. That is the first thing I remember clearly. It was aggressively, almost offensively warm.

I had not realised how cold I was until I entered the building and my hands began to hurt differently, with the sharp pain of circulation returning.

They treated me for hypothermia, which was not a surprise. They cleaned and dressed the lacerations and documented them as consistent with contact with barbed wire.

I did not correct the clinical notes. I did not know what explanation I could offer instead.

I slept for several hours.

When I woke, an Environment Agency investigating officer was sitting in the chair beside the window with a notepad. She was perhaps fifty, wore a grey jacket and had reading glasses pushed onto the top of her head. She introduced herself, but I have since forgotten her name.

She was calm and said she was sorry for what had happened. Then she asked me to take her through the survey from the beginning.

I told her everything.

I described the earlier GPS anomalies, my replacement GeoNav XR-40 receivers and the notebook. I explained that the real Elliot's contractor paperwork was filled in with rounded, right-slanting handwriting, while the entries written during the survey were cramped and slanted to the left. I told her about Professor Peter Wycliffe, the name the false Elliot had given me, and that Wycliffe had retired in 2020 and died in 2022.

I described the figure standing on the water, its silent descent without bubbles and its later appearance beside the derelict shooting hide. I told her what I had seen happen to Gareth, the flare pistol and the burning hide. I told her that Gareth had gone under and had not resurfaced. I told her about Silas Fenwick's punt, with its red transom and patched port gunwale, waiting on the bank.

She wrote down everything.

She said my statement would form part of a joint investigation. Given my role as lead ecologist, there would also be a formal internal review. She told me I was suspended from field duties pending a joint Environment Agency, Broads Authority and police investigation into the disappearance of a ranger and a university contractor during a routine wildlife survey.

She said 'routine' as if apologising for the word.

Routine.

The word did something small and unpleasant inside me.

I signed the paperwork.

They returned my gear bag, which Daniel had recovered from Silas's punt before the ambulance departed. I checked my mobile. The photographs were still there.

There were 12 images, covering every relevant page of the field notebook. Each showed the cramped, left-slanting entries. The writing bore no resemblance to the rounded, right-slanting hand in Elliot Price's contractor file, which I had reviewed two weeks earlier in Norwich.

The audio file was there as well.

Thirty seconds. I locked myself in the hospital lavatory, put in one earbud and listened to it once. The humming repeated in an exact four-second loop, too even and regular to be absent-minded or accidental.

I removed the earbud and returned the mobile to my pocket.

Both replacement GPS receivers were in the gear bag. They were the new GeoNav XR-40 units I had ordered after dismissing the earlier teams' readings as calibration faults.

While I waited for my discharge papers, I switched one on.

It took about forty seconds to acquire a signal. Then it displayed the same coordinates in Hickling Broad and placed me 9.4 m beneath the water surface.

I turned it off and put it back in the bag.

They discharged me on Monday afternoon, 16 November 2026.

I drove back to Potter Heigham in my vehicle, which Daniel had arranged to be moved from the slipway to the hospital car park. I do not know who drove it. The seat had been adjusted slightly too far forwards.

The Potter Heigham village sign stood beside the road as I returned. I had passed it on Saturday on my way to the survey and scarcely noticed it. Now even the name seemed to belong to a place that was uncertain of its own borders.

My accommodation was a small rented cottage on a shingle lane outside the village, not far from the water. It had two rooms, a temperamental bottled-gas cooker that usually took three attempts to light and a kitchen window overlooking reeds and willow trees.

I had rented it for a fortnight because the survey programme was meant to last a fortnight. Twelve days remained on the booking.

I placed my gear bag on the kitchen table and sat beside it.

The 12 photographs were on my mobile. So was the 30-second audio recording with its four-second loop. My own written coordinates remained in my waterproof notebook. The archived records from both earlier teams gave the same coordinates, and all three sets of receiver logs recorded the same impossible elevation: 9.4 m beneath the water surface.

The old drainage-board and waterways survey chart held in the Environment Agency archive marked an 11.3 m hollow at that exact location. Beside the 1957 drainage-board survey entry was a recommendation for a repeat survey that had never been resolved.

Three separate teams. Three sets of data reporting the same thing.

I had not believed the first two. I had written 'probable calibration fault' in the margin and ordered replacement receivers.

I thought about Gareth Morgan, the level cadence of his voice and the way he had spoken about Silas Fenwick at the slipway as though telling a familiar local story. Gareth had remained calm throughout the day, even after events stopped making sense. When he saw the figure, he moved towards it because he was the sort of person who moved towards problems.

I had been only metres away, and I had not been fast enough.

I thought about Elliot Price, who had driven to the slipway before dawn but disappeared somewhere between his Hilux and the water.

The investigation would examine my field reports, equipment logs, risk assessment and every decision I had made during the operation. They would review the receiver data, the 12 photographs and the audio recording with its exact four-second pattern.

Eventually, after meetings, paperwork and repeated use of the word 'routine', someone would have to return to that channel.

I knew that. I had documented enough to make another survey unavoidable.

The channel was still there. Whatever had taken Gareth had not burnt. The flare had ignited the hide, and the hide had collapsed into the water.

Afterwards, the thing had simply ceased to be visible.

I believed it might have returned to the 11.3 m hollow where it had waited before we arrived, but belief was not evidence. I could not prove that it was there, only that neither it nor Gareth had been visible when I left.

It could be waiting there when the next team launches from the Broads Authority slipway in the rapidly failing November light.

That team will go because of my records. Because I dismissed two warnings as faulty equipment, then followed procedure until procedure could no longer keep anyone alive. Because the evidence I preserved is too consistent to ignore.

I switched on the GeoNav XR-40 once more and watched it acquire a signal.

The same coordinates appeared.

The receiver placed me 9.4 m beneath the water surface.

I left it running on the kitchen table and went to bed.