We study flash floods, flash droughts, heatwaves, extreme precipitation, compound events and extreme winds, and we measure what they do to agriculture, energy, biodiversity, food security, economies, insurance and public health. We work for governments, enterprises, companies and small businesses anywhere in the world, online or on your site. Every job ends with a number, the error margin around it, the code behind it, and a recommendation you can act on.
We study six kinds of extreme event and measure what they do to seven things people depend on. Every study ends with recommendations, not just numbers.
Rain arriving faster than the ground or the drains can take it. How intense the rainfall gets over one, three and six hours, how often a design threshold is crossed, and how little warning there is. Flood mapping built on daily totals misses these entirely.
The rainfall itself rather than what it causes. Intensity, duration and frequency curves, the wettest day and wettest five days of each year, design storm totals, and whether any of those are rising. These are the numbers engineers size drainage, culverts and spillways against.
The fast ones. Soil moisture collapsing over weeks rather than months, driven by heat, wind and no rain together. Monthly drought monitoring misses them completely, which is why they catch people out.
How hot, how long, and how warm the nights stay. Humid heat, which is what actually harms people and livestock, and how often the thresholds that matter to crops, workers and equipment get crossed.
Tropical cyclones and hurricanes, tornadoes and squall lines. Peak gust return levels, how often a design wind speed is exceeded, and what that means for turbines, transmission lines, roofs and standing crops.
Two hazards in the same place at once, or one straight after another. Heat during a drought. Flood on ground already saturated. Treat them as independent and you will underestimate how often both happen, usually by a factor of four or more.
When the rains start and stop, how long dry spells run, heat during flowering, and drought and heat landing in the same season. All of it mapped onto planting windows, input decisions and expected yield.
How much wind and sun a site really gets and whether that is changing, how much a plant will produce next week, what heat does to panel and transformer output, drought risk to hydropower, and demand peaks during heatwaves.
Vegetation stress measured from satellite records going back decades, how often a protected area now crosses heat and drought thresholds it used to see rarely, habitat exposure to shifting rainfall, and fire weather.
The path from a rainfall deficit to people not being able to afford food. Production shortfall, how markets respond, which districts are hit first, and how far ahead the warning signs are visible.
How extremes travel through prices, trade and output. Shocks to major producing regions, how a harvest failure moves a commodity price, and what sector output looks like under warming scenarios.
How severe a one in fifty year event is and how wide the error margin around that is, how often two perils strike together, where a parametric trigger should sit, and how often it will miss a real loss.
Heat and mortality, particularly among outdoor workers and the elderly. How warming and shifting rainfall move the season and the range of malaria and dengue. Cholera and diarrhoeal disease after floods. When and where a health system should expect the load.
If your exposure to these hazards sits outside the seven sectors, tell us what it is. If we cannot answer it well we will say so on the first call, and there is no charge for finding out.
The analysis itself. How often the event happens now, how severe it gets, how wide the error margin is, how much of that is changing and how much is natural variation. You get the written report, the charts as separate files, the data we built, and the code that produced all of it.
What to do about it. Options ranked by what they cost and what they avoid, written for the people who will actually decide. For government work that means something a committee can adopt and defend in public. For a company it means a decision memo your board can act on this quarter.
Every one of these can be delivered online, or on your site anywhere in the world. Most clients choose online. We travel when the work needs us in the room or in the field, and we quote the travel before we start.
These are the packaged versions of the work above. Four have a fixed price and a fixed delivery date and you can book them today. The larger pieces need a conversation first, because the price depends on what your data looks like. All of them run online or on your site.
Ministries, meteorological services, disaster agencies, provincial and city governments. Work that has to stand up in public and feed into a policy decision.
Humanitarian organisations, development programmes, foundations and UN agencies. Triggers for anticipatory action, evidence for a funding proposal, and targeting decisions taken before a season turns rather than after.
Universities, research institutes, graduate schools and funded consortia. We join projects as a research partner, teach the methods, and co-supervise the students doing the work.
Insurers, utilities, banks, agribusiness groups. Board-level risk work, usually under a confidentiality agreement and often run on your own data alongside ours.
Mid-sized firms with one specific exposure. One site, one crop, one season, one decision that needs a defensible number behind it.
Farms, cooperatives, small operators and consultancies. You need one clear answer rather than a research project, scoped so it costs less than the decision is worth.
One question, one place, a straight answer. For anyone who needs a defensible number this week and does not need a full study to get it.
We look at one country or growing region for the season ahead. You get a short report you can take straight into a planning meeting.
If someone will ask you to defend a number, this is that number. We also tell you how wide the margin of error is.
Two hazards often arrive together. Drought and heat, for example. If you treat them as independent, you will underestimate how often both happen, usually by a factor of four or more.
We pick the index, set the payout levels, and measure how often the trigger will miss a real loss. For insurers, brokers and government risk pools. Six to eight weeks.
How much wind and sun a site really gets, how much a plant will produce next week, and what that means for the grid. Built for tropical conditions. Six weeks.
Adaptation options ranked by cost and benefit, written for a ministry, a regulator or a development bank submission. Eight weeks.
We join your project as a partner on the climate extremes work package. Proposal writing, the analysis itself, and co-authorship. Often funded directly from the grant rather than paid by you.
Graduate courses and intensive schools on extreme value statistics, compound events, and handling climate data in Python and R. Taught on real datasets. Your students leave with code that runs.
Co-supervision of MSc and PhD projects on climate extremes and their impacts. Weekly contact with the student, methods guidance, and review through to submission. We have supervised six MSc theses to completion, five with distinction.
Each month you get a briefing on your markets and two calls with us. Small questions get answered the same week. You also get first claim on our time when you need a bigger piece of work. This suits insurers, agribusinesses and energy firms who need this every month rather than once.
Start a retainerTwo days in Python, taught on your own data. We cover extreme events, hazards that arrive together, and how to handle climate datasets without making the usual mistakes. Minimum eight people. Your team leaves with code that runs.
Book trainingPrices do not include VAT. EU businesses with a VAT number are invoiced under reverse charge. Clients outside the EU are invoiced without VAT. Fixed-price work is paid when you book. Retainers are billed on the first of each month.
Climate extremes are our specialism. This is the second thing we sell, and it is the same people and the same tools pointed at a different subject. Most organisations we meet are sitting on data they cannot use. It is spread across forty spreadsheets, nobody agrees which version is current, and the monthly report takes three days to assemble by hand. We fix that.
Send us what you have. We tell you what state it is in and what it can honestly answer. Start here if you are not sure the rest is worth paying for.
We move your data out of spreadsheets into one PostgreSQL database, then build the reports so they run themselves.
Sales, demand, prices, yields, energy use, staffing. If you have a few years of history, we can usually forecast it.
Pick a task someone does by hand every week. Reading reports, pulling numbers together, drafting the same summary. We build an agent that does it, with the same checking step we use on our own work.
Moving off Excel, Access or an ageing MySQL setup onto PostgreSQL, without losing history and without a week of downtime.
Someone to call when a query breaks, a report looks wrong, or you need a new number before Friday. Includes small changes and one call a month.
Where your data lives. One database everyone reads from, instead of forty spreadsheets emailed around with different numbers in them.
The analysis itself. Cleaning, statistics, machine learning, and the scripts that produce your reports while you sleep.
Statistics and charts, for work that a statistician or a regulator will check line by line. Some clients need this and we do not argue.
The repetitive parts. Reading long documents, writing the first draft of the code, running the same job every month. We check everything they produce.
Same terms as our climate work. Prices do not include VAT. Fixed-price work is paid when you book. Our two-day training course also covers Python, R and SQL on your own data, at €480 a seat.
Thirty minutes, free. You tell us the decision and the deadline. We tell you whether the data exists to answer it and which service fits. If none of them fits, we say so.
We send a quote with the price, the scope and our bank details. You pay by transfer. Larger jobs run on a signed proposal with payments in stages.
Online, or on your site where the job needs us there. Halfway through we send you what we have, so you can change direction while there is still time.
The report, the charts, the code, and a call to go through it. It is all yours. No licence fee, no per-seat charge, nothing to renew.
Most of our work. Analysis, modelling, code, reports, review and advisory calls, delivered from Germany and our African offices to clients in any time zone. We hold calls in your working hours, not ours. Nothing about climate data analysis requires anyone to be in the same room.
Some things go better in person. Training your team, working alongside your analysts through a model build, collecting or checking field data, inspecting an instrument network, or presenting to a board or a regulator. We travel anywhere. Travel and accommodation are quoted before we start and never added afterwards.
This is why an assessment takes four weeks and not three months. AI agents can read a whole literature, write the data code and draft a report in a day. The problem is that when they get something wrong, the output still looks right. So we check everything they produce before it reaches you.
AI labs pay our founder to break their own models and write down how they fail. That is an odd thing for a climate consultancy to have on staff. It is also why we trust AI with your work. These are the four mistakes we look for every time.
Every recommendation we give you comes with the number, how certain we are about it, and a note on what the evidence does not tell you. Most consultants leave out that last part.
Photon is a new company. The science behind it is not. These papers went through peer review and anyone can read them.
We started Photon because the people who need these numbers most rarely get them in a form they can use. An insurer wants a figure with an error margin on it. A ministry wants a planting advisory it can publish next week. A grid operator wants to know about tomorrow, not about 2050. Universities are not set up to work to those deadlines. We are.
We keep the work narrow on purpose. Climate extremes and what they do to agriculture, energy, biodiversity, food security, economies, insurance and public health. We use extreme value statistics, machine learning and AI agents. We do not write sustainability reports or strategy decks. We give you the number, the code behind it, the options it points to, and a note on what it does not tell you.
We work with the African Climate and Energy Research Group, a network of climate, energy and environmental researchers across the continent. They cover renewable energy, climate patterns, climate impacts and atmospheric science. It is how we reach local knowledge and local data in days instead of months.
Our offices are where our own people sit. The work itself goes anywhere. Climate data analysis does not care where the analyst is, and when a job needs us physically present we travel. Clients in Lima, Jakarta or Warsaw get the same service as clients in Nairobi.
Our registered office and where the analysis is done. Darmstadt sits next to the European weather and satellite centres, which is useful when we need data fast. All contracts and invoices come from here.
Our Central Africa base, at the foot of Mount Cameroon. Coastal and forest zone work, and our oldest links to the University of Buea and AIMS Cameroon.
West African coast. Cocoa, coastal flooding, and the harbour and power infrastructure that sits a few metres above sea level.
The Sahel. When the rains arrive and when they fail, groundnut and millet yields, and the early warning systems that depend on both.
East Africa and the continent's biggest market for weather insurance. Drought cover for pastoralists, tea and horticulture, and the insurers who underwrite them.
The Great Lakes and the highlands. Steep terrain, intense rainfall and landslide risk, where a coarse climate model tells you almost nothing.
Southern Africa's maize belt and its power system. Drought and heat over the Free State and Mpumalanga, and wind and solar forecasting for the grid.
The dry southwest. Water supply, rangeland and livestock, in a country where a bad season is measured in boreholes rather than in yield per hectare.
Behind the African offices sits the African Climate and Energy Research Group. When a job needs someone who knows a specific river basin or a specific crop, that is where we find them.
Outside these eight countries we have no office and we do not claim one. We do not need one. The datasets we work with are global, so an analysis for a client in Peru or Indonesia runs on the same archives and the same methods as one for Kenya. Where a job needs us on the ground, we come to you and the travel is in the quote. Where it does not, you will never notice the distance.
Our fixed-price work starts at €1,500 for a single hazard at a single site. A full return-level study for one location is €2,400, a compound-event study is €4,500, and a multi-site portfolio assessment is €8,500. Larger or unusual jobs are scoped after a free call. Every price is agreed in writing before any work starts.
It means a flood with a one per cent chance of happening in any given year, not a flood that arrives once a century. Two of them can land in the same decade without anything being wrong with the statistics. We fit the extreme tail of your local record, report the return level you asked for, and give you the error margin around it, because that margin is usually wider than people expect and it changes what you should build.
It is two or more hazards that hit together or in sequence, where the joint damage is worse than the sum of the parts. A heatwave during a drought. Extreme rain on ground already saturated. Low wind and high demand on the same cold night. Treating each hazard separately underestimates the risk, sometimes by a factor of several, because the hazards are statistically dependent. We measure that dependence with copulas and report joint return periods.
A drought that develops in weeks rather than seasons, driven by heat and high evaporative demand instead of a long rainfall deficit. It is dangerous because it arrives faster than most monitoring systems and most irrigation decisions can react. Conventional drought indices that work on monthly rainfall often miss it entirely.
Yes. We work anywhere in the world, online or on your site. Our head office is in Darmstadt, Germany, and we have offices in Ghana, Cameroon, Kenya, Senegal, Rwanda, South Africa and Namibia. A large part of our published work is on African climate, and we handle sparse station data and short records, which is the usual problem outside Europe and North America.
Often nothing. We can work entirely from public reanalysis, satellite and model data, which covers the whole world. If you have your own station records, yield data, loss records or outage logs, the study gets sharper, because we can tie the hazard numbers to what it actually cost you. We will tell you on the free call which of the two applies to your case.
A single-hazard briefing takes about one week. A return-level study takes two to three weeks. A compound-event study takes three to five weeks. A portfolio assessment takes six to eight weeks. We give you a date in the quote and we tell you early if anything is going to slip.
Yes. Physical risk disclosure requires hazard numbers for named locations under named scenarios, with the method written down. That is exactly what we produce. We give you the figures, the scenario assumptions, the uncertainty, and the code, so your auditor can see how each number was reached. We are not auditors and we do not sign off your report, but we supply the evidence it rests on.
Yes, with every job. You get the scripts, the processed data and the notebooks that produced each figure. Nothing is locked inside a tool you have to keep paying for. If you hire us once and then want to run the analysis yourself next year, you can.
AI reads, drafts and searches. It does not decide. Our agents go through the literature, pull candidate datasets and draft the first version of the write-up, which saves weeks. Every number that reaches you comes out of statistical code we wrote, and a human checks the result against the raw data before it leaves. If a figure cannot be reproduced from the code, it does not go in the report.
The more you fill in, the more useful our first call will be. Only the starred fields are required. Send it half finished if that is all you have time for.
We will reply within one working day, usually sooner.
Your answers are below. Copy them into an email to us and we will take it from there.
ayuketang@aims-cameroon.orgYou tell us what you need, we send a quote with the price, the scope and our bank details, and you pay by transfer. No card details are taken on this site, and we do not store any.
The quote we send carries our full bank details and a reference number. Payment within 14 days for fixed-price work, 30 days for institutions.
EU businesses with a VAT number are invoiced under reverse charge. Clients outside the EU are invoiced without VAT. We can invoice in euros, US dollars or pounds.
Coming shortly. If paying by card or mobile money is the only way your organisation can buy, tell us in the form and we will arrange it.
ayuketang@aims-cameroon.org+49 157 3380 8636