Dr. Yumi Saito had been a geologist for fifteen years and a catastrophist for three.
She made this distinction herself, with the particular clarity of someone who has learned that precision in self-description is the first step toward being taken seriously in rooms where people do not want to hear what you have to say. A geologist studies the earth. A catastrophist studies what the earth does when it has been pushed past its tolerance — studies the edge cases, the tail events, the scenarios that appear in probability tables under columns labeled UNLIKELY and are not, in the geological record, particularly unlikely at all. They just don't happen on human timescales. Until they do.
She told Kenji all of this in the first two minutes of their conversation, while they were still standing at opposite ends of the concourse, gradually closing the distance between them in the way that two people who are both trying to assess each other approach each other: cautiously, directly, without pretending the caution isn't there.
He listened with the attention she recognized from other engineers — not the passive listening of someone waiting to speak, but the active listening of someone building a model from the incoming data, checking it against what they already know, flagging inconsistencies.
"Your models predicted a magnitude 8.4," she said. "Single event, shallow focus, Koto epicenter."
"Yes."
"My models predicted the same event. Same magnitude, same location, same timeline."
"But you said there's something I missed."
"Yes." She opened the red notebook. The pages were dense with data: hand-plotted seismic frequency maps, chemical composition analyses, something he'd only seen in deep geological research, not in seismic prediction. "The rock chemistry under the bay."
"I reviewed the bay surveys."
"You reviewed the structural surveys. I reviewed the geochemical surveys, which are different documents from a different department, and they're not usually cross-referenced with seismic prediction because they're considered separate domains." She turned the notebook to show him a graph. "There are elevated levels of methane hydrate destabilization in the sediment below the fault zone. The same destabilization pattern that preceded the 1995 Kobe earthquake and the 2011 Tohoku event, but larger."
Kenji looked at the graph. He looked at it for a long time.
"Methane hydrate destabilization at that depth, in that concentration —" He stopped.
"Would significantly amplify the seismic event," Dr. Saito confirmed. "Not the primary wave — the secondary wave. The one that comes six to twelve hours after the main event. Your models predicted a seventy-three-percent probability of a secondary event. Mine predict a near-certainty. And not at 7.0 to 7.5, which is what you would have modeled."
"At what?"
She said the number.
Kenji sat down. Not on a bench — just on the floor of Tokyo Station, with his back against a column, because the number was large enough to require sitting and there wasn't a bench close enough.
"That changes the casualty projections entirely," he said.
"Yes."
"The buildings I assessed as moderately resilient — the structures I considered recoverable —"
"Many of them would not survive a secondary event at that magnitude."
"How many people —"
"I don't want to give you a number in this context," Dr. Saito said, carefully. "Because whatever number I give you will prevent you from thinking about anything else, and right now we need to think about what this means for your system."
She sat down across from him on the marble floor, cross-legged, red notebook open on her knees. In the golden light of the empty station, she had the composure of someone who has been carrying large numbers in her head for months and has learned to be in the same room as them without being consumed.
"Your system takes families to a safe parallel timeline," she said.
"Yes."
"How many families does it currently have capacity for?"
He thought of the screen on the Shinkansen. Eleven other families currently in trial.
"It's running eleven families simultaneously. It was designed for one."
"And the anchor — the version of you holding it together — is under strain."
"Yes."
"Mr. Tanaka." She held his gaze steadily. "With the revised secondary event projections, the number of families that need to be moved before the earthquake is not eleven. It's substantially more. And the window — your timeline has already accelerated once. Based on the geochemical data —"
"It may accelerate again," he said.
"Yes."
They sat with this together on the marble floor, two engineers who had arrived at the same catastrophe from different directions, in an impossible space between worlds.
"The system needs to be expanded again," Kenji said.
"Or a second system needs to be built."
"The anchor can't support further expansion. I know that."
"Then a second anchor is needed."
"The system requires —" He stopped. Thought about the prompt on his laptop screen three weeks ago. THREE OF THE FOUR PORTALS ARE READY. THE FOURTH REQUIRES A SACRIFICE. "The anchor position requires someone to remain in the between-space. Indefinitely. Running the system from inside."
"I know," Dr. Saito said.
"You know."
"I know because your notebook was very complete. And because I've had fourteen months to think about it." A pause. "And because I have no family in this system. No one waiting for me on the other side. My parents died young. I was never —" She looked at the marble floor. "I was never very good at the things that require you to stop working long enough to build them."
Kenji looked at this woman he had never met, this geologist who had found his notebook and followed his trail and come through an accidental portal into his trial environment, and he understood what she was saying.
"I can't ask you to —"
"You're not asking. I'm telling." She closed the red notebook. "I came through this portal because I was trying to reach you. Because I thought that if I could get your data and my data into the same room, we could figure out how to help more people. That's still what I'm trying to do. The fact that the way to do it is this —" She gestured, loosely, at the empty station around them, at the whole impossible situation. "— doesn't change the calculation."
"It changes everything about the personal cost," Kenji said.
"To me, or to you?"
He had no answer for that. Or rather, he had an answer and it was not the one he wanted to give, which was: to both of us, and to the part of me that is already holding this weight alone and knows what it does to a person.
Dr. Saito stood up, dusted off her field trousers with the practical manner of someone returning to work. "Your trial isn't complete," she said. "The station is waiting for something from you. What is it?"
Kenji looked at the departure boards. The footage of his Tokyo was still running — the Sumida River, his other self on the bank, the city going about its ordinary business above an extraordinary threat.
"I need to say something I haven't said to anyone," he said. "Something true."
"Then say it."
— ✦ —
He looked at the footage. He looked at the man on the riverbank — the version of himself who hadn't come through, who would face what was coming without the system, without the portal, without any of the preparation Kenji had spent three years building. He thought: that man is going to try to do something. He won't give up. He'll find another way, or die trying, and either way he'll do it alone because that is what we do. That is what I taught both of us to do. He opened his mouth to speak. The true thing. The hardest thing. And from the departure boards, from the footage of the Sumida River, the other-Kenji — the one on the bank, the one who should not be able to know he was being watched — raised one hand. A slow, deliberate wave. Acknowledging him. And then the man began to move, and Kenji realized: he wasn't waving. He was pointing. Pointing at something behind Kenji, something in the station that had not been there when he'd arrived. He turned. There was a door. Not the wooden door of his trial entrance. A steel door, utilitarian, with a keypad and a sign he couldn't read from this distance. The door was open. And through it, the sound of a city in motion — not the golden silence of the empty station, but the full, complex noise of Tokyo at full operational capacity: trains and crowds and the indifferent roar of fourteen million people going about their lives. The door to the origin timeline was open. And standing in it, blinking at the impossible golden light of the trial station with the expression of someone who has broken through a wall and is not quite sure what they're standing in, was a girl of perhaps fourteen. She looked at Kenji. She looked at Dr. Saito. She looked around at the empty station with the fast, cataloguing eyes of someone who is scared and covering it with assessment. "I found the frequency," she said. "I've been tracking it for six months. I thought —" She swallowed. "I thought there had to be a way through. My whole neighborhood. Everyone I know. Someone has to help them get out."