Why does this exist?

Because technical advice in the field improves when calculation, data and diagnosis stop being the bottleneck, and that is the optimistic view, because most of the time those three are never even seen together. Much of what we know stays in a paper that 95% of people cannot even reach, in the manual or in somebody's spreadsheet, and never reaches the daily work of the person who has to decide.

Everything here came out of the field and goes back to the field: calculations that used to be done on a loose sheet, diagnoses that depended on the eye alone, and data that got lost in a notebook. The idea is to make what already exists accessible so that each day's work is a little better.

None. There is no third-party analytics and none of our own either: no Google Analytics, no self-hosted alternative, no pixels, no heat maps. No advertising, no user profiles and no selling of data. Just get on and enjoy the tools.

The only thing there is is the account panel: your registration and whatever you choose to save in the tools, which is exactly what is needed for this to work. Plus technical cookies: keeping the session, protecting the site and remembering preferences such as dark mode. There is less tracking here than in a secret agency, and it is all set out in the privacy policy.

Because there are now fewer real people on the internet than machines. Sad but true.

All the tools are complex code processed internally on a dedicated server. This is not the usual site you find on Google, the kind that only wants you in to grab your data and hand you a poor tool. It is safer that way, but it also costs: every calculation is heavy and consumes processing capacity. And there are bots trying to use them all day long, at my expense. Registration is the simplest way I have found to tell someone who works in the field from a robot that just wants to burn CPU.

Yes. I would have liked something like this when I started, instead of spending hours, years, only to find out that plenty of things are not even done with any real criterion. Advisers' work needs making easier, and we all have to contribute something for this to improve, otherwise we are back to everyone looking after their own skin. The basic tools are free and will stay that way: if they were paid, I could never have afforded them when I started, and not everyone has the same resources or lives in the same place.

Yes, and it is genuinely appreciated. The server is dedicated, the calculations are heavy and it gets paid every month (an annual contract would be better, but I cannot afford one). Prices, with the memory and chip shortage that AI has brought, are at levels never seen before. It all goes to the same place: keeping the free tools standing and building the ones still missing.

That said, only if you can afford it. And it changes nothing of what you see. There is no paid version, no extra access for contributing, and no record of who put in what. The same tools, free for everyone, whether you contribute or not.

It depends on the group and the access level of your account. I try to keep the basic ones, which you will not find elsewhere even paying and will not find at all, completely free, and they will stay that way, along with the new ones coming out of development and improvements.

The growers and companies I work with have access to advanced tools and specific services, because those are the ones I use with them on their farm. That is the perk of having me working in your company.

No. The tools I keep aside are for my own use or for the companies and colleagues I work with. Some are very specific. Others, like the genetics or experimental design ones, are for researchers, because the commercial versions are glorified spreadsheets, and to use those you are better off with your own spreadsheets. The online donations on the home page help me put more time into it, finish the planned ones and add more complex ones as well.

For faults you spot, yes, and for all users: if something breaks, report it and it gets fixed. I cannot answer hundreds of emails asking why the acid or the nitrogen gives that figure.

Agronomic training, on the tools and on how to apply them to real work, goes with the on-farm advisory work and with how much you want to learn, not with the site.

Quite varied, all agronomic and at professional level. Fertigation and plant nutrition (nutrient solutions, inverse calculation by target EC, liquid fertilisers, fertiliser tank balances, injection calibration), advanced nutritional diagnosis, irrigation water analysis and water blending, soil analysis and interpretation, organic and broadacre fertilisation.

Also plant health with degree-days and pest phenology over commercial and official weather station data, plant protection product lookup, ETo and irrigation design, planting density, the farm logbook, phenotyping and trials with design and full statistical analysis with an API, and two in-house GIS (web and mobile) integrated with SIGPAC land parcels.

I do not take custom requests. The tools here exist because they are useful, because they are not available anywhere else and because they do a job. Even so, suggestions and ideas are welcome: if one is workable, I may build it some day.

No, they are different, although both are a full GIS. The mobile GIS is a separate application, with its own code for performance and features as well as for agronomic and management modules, which tries to cover the technical side a technician needs in the field, and it works offline too: photos, analyses and the advisory work in one place. The GIS on the site is integrated with the site data and gives access to many tools, custom layers and geometric processing. It is not a viewer: it is a real GIS, customised and modified by me.

Because the commercial alternatives are absurd: they offer little at the price of a lot, and honestly none of them measures up to what a technician needs in a company that is already stretched. Living off forms is worse than having nothing, because without them you are at least more productive. Today you need many tools at once, and mobile GIS apps sadly stay as prototypes behind a commercial mask and SaaS services. That is why I have spent years building my own.

To give you an idea, it has more than twice the code of the best GIS you can find on a phone today. It works offline, there are no subscriptions and it runs on its own, but it keeps compatibility with QGIS, so you can work in either one and move things across. The engine is optimised in C++ and handles layer threads asynchronously, which none of the others do, and that is another matter: it allows data and remote sensing levels far, far beyond what institutions and many companies offer, and with hundreds of thousands of features it performs better than desktop QGIS.

From published literature, from regulations and official data, and from what I measure myself. You often come across monumental inconsistencies in data, dry matter for example: it is almost comical how studies at the same temperature show differences of more than 300%, even in the same organs, and there is no way to work like that. Uptake figures and the split between organs come from specific studies, averages of my own, with their crop and their conditions, not from a round average going around the internet, and not from work assumed to apply when it does not and that nobody can trace back.

Every tool shows the assumptions it works with. Over the years I have developed much of it myself from studies, and that part is either not cited or cited generically, because in the end a lot of it needs a local or applied approach. When there is no figure for a crop, it is left empty and said so. If something does not match your area or your variety, write to me with the figure and the reference: it gets checked and corrected. In the tools I normally let you change almost everything, because this is for technicians, not for someone on a guided tour.

Because this is for technicians. There is no customer here to keep happy with the easiest possible app so that they pay. It is for advisers who know what they are doing, and there are management situations where you can do what you would not do in others.

With the water. An irrigation water analysis properly understood, then the soil or the substrate, then the nutrient solutions. Most of the fertigation problems I see in the field come from water read badly and used badly, not from a formula worked out badly, because all that does is propagate the errors. The moment you realise that the calculations for a fertilisation plan are much more than, and very different from, uptake figures, and in many cases make no sense, you are on the right track.

And with your own data in front of you, not with exercises. Those help, but in the end turning what you see into something you can handle is very powerful and makes learning much easier. Take an analysis you already have, run it through the tool, look at the assumptions it shows and compare them with what you were doing. That is where you learn, and contrary to what is claimed by those who sell training rather than knowledge, what is worth learning is hard. The best way is when you use something you do not know and have to stop and ask yourself why, when the figure does not match what you expected and you have to work out where it comes from.

Spanish and English.

Ismael Maldonado, chartered agricultural engineer. I work in the field every day doing integrated advisory work and agricultural engineering consultancy in the intensive horticulture of Almería, specialising in fertigation and plant nutrition, with an in-house laboratory to diagnose on an analytical basis. Yes, you read that right: I am not going to wait weeks for a laboratory to diagnose something I can diagnose the same day. The grower is happy and the company gains a service.

This is not a department's project, and there is no funding behind it: it is the work of one person who advises, works out how much time is lost repeating the same thing and then turns it into a tool. The full background is on About.

From the start I saw a pressing need in my work: knowing whether this is one pest or another, applying the right fungicide at that moment and not 3 weeks later or blind, and cutting preventive treatments that are a cost and a problem for the crop and the environment.

Knowing deficiencies, knowing whether there has been internal cavitation that has damaged the plant's vascular bundles, seeing the effect of products on organisms that are often resistant, or the real behaviour of entomopathogenic fungi depending on humidity, residues and light spectrum, which brings surprises between what you are told and what you see. For that you need the equipment, the knowledge that builds up over time and the will. For me it creates a need, and that has a direct impact on the quality of the technical service. At the end of the day I am the professional who has to give a good service. That is why I see it as an integral part of my work, included with me on the payroll, even though it is mine. I do not see how this can be replaced for informed, sustainable decisions, but maybe that is just me.

You hear it more from technicians than from growers, which already says how serious this is. Often it is simply a symptom of the underlying problem in agriculture: knowledge is local and has to be adapted to the production system. That often means that moving 200 km puts you in a completely different production system, so what is taught and what is seen does not apply the same way. That breeds complexity, weariness and in the end the feeling that, if you apply x, y and z from the mix, you are better off doing nothing. That is said by whoever applies the same thing every year and calls it experience. Calculating is not filling in a table: it is knowing how much goes in, how much the crop takes away and what is left in the soil, and when you should, or it pays, to stay on the line or go to one side or the other. Whoever does not calculate is deciding too, only by eye and with no way to explain why.

And then there is the bill. Five per cent of yield is easily lost by applying a fungicide three times the wrong way. Ten per cent of yield is normally gained with good nutrition and good crop control, and in a greenhouse that is a lot of money. Contrary to what is said all the time about labour, inputs have doubled and tripled in recent years: they are not a small line, they are one of the biggest costs. Too much is money thrown away, salts building up and nitrates ending up where they should not. Too little is yield and quality you do not get back. It takes less time to work it out than to argue about whether it is needed. Imagine you are a grower or a company and the technician you hire tells you up front: «you are going to lose ten per cent of your yield on average, because I do not intend to manage the crop's nutrition». What a great start, no doubt.

None of them. In fact, if you ask me, I no longer remember what each one was. In a competitive world it is fine for winning share, but here I have no share to win. If you take into account how the pace of adopting technology has slowed, had it kept the pace of years ago we would be somewhere else, but no: much of what new equipment does today I can do with twenty-year-old equipment, literally, the sort still sitting on farms. The outlook has to change. Technology has to be affordable, but also useful and with an impact. Without that impact it is one more cost. If you make me put a label on it, pull on two threads and you end up back with the agricultural researchers of 1960 to 2000, since we are still doing much the same. This is still the agronomy of always with the means of now: measure, calculate, decide and leave it recorded.

What changes over the years are the means, not the criteria. A badly framed balance does not improve because the figure arrives from a sensor: measuring badly leads you to worse decisions than not measuring at all. Judge a tool by what it calculates and by the assumptions it shows you, not by the version number it claims.

The future is management. Technology helps when it pays its way and gets in the way when it only generates costs, is badly applied or keeps you clicking. A parral greenhouse, with almost no ventilation and uncontrolled water, is not fixed by filling it with sensors: if you do not adapt the measurements and the management, it stays the same, but with charts.

What does change things is measuring where it matters and acting in time, with the figure next to the decision and not in a report three weeks later. First what holds the crop up, climate, water, soil and plant health, and then what measures it.

Because when a single path was imposed, years ago, many sound practices were left behind, and the judgement to apply them went with them. This gets seen as a business rather than as a production model, and that is where the atrophy and the lack of interest come from. Today there is more information, and for many of those practices we already know whether they hold up or not. Something being traditional does not make it good: what does not add up is going back to the yields of then with the costs of now. That is why new techniques and management that bring the two together matter, because sustainability has to be profitable. Otherwise we wreck the sector.

Not the agronomy. In fact most of the tools existed in a more primitive version years ago, before these language model AIs. The criteria, the equations, the references, the values and the algorithms behind every tool come from my own work in the field and from the literature I use, and I check them one by one against real cases before they go in. A model has never stood in front of an irrigation head at seven in the morning, and it does not know where an uptake figure comes from or why that figure is no good for the crop next door.

For writing code I do use it, mostly for reviewing, for syntax and for repetitive work that eats your time. A site this size with an interface does not get done by hand: the alternative would be keeping the raw calculations in the scripts and programs I had, which are frightening. It gives me what I always say about accessibility: an organised way to offer what I already had and already do. AI also has enormous drawbacks. Its knowledge is purely what it believes a paper says, and what it gives you back when you ask is that paper. In many cases this ends in the generic being applied, and we go back to the old days of applying things literally, which does not work in agriculture because systems vary so much. Not to mention that, if you feed it things that contradict each other, an AI never wants to say «I do not know», and that is its biggest failing. All that speeds up the mechanical part. It does not decide the agronomy, and it should not.

No, none of them uses it to calculate. It is not needed: as things stand, AI is the wrong tool for almost all of this. That will change with time, but the real problem is the lack of useful algorithms applied to agriculture (nothing new there), and that is not solved with an AI but by modelling, running trials and checking. Without experience behind it, it cannot be done, however much AI helps you.

Probably a case of the cobbler's children going barefoot. I know how what we now call AI works underneath (the basics, at least), which is really language models. I have a customised inference engine so I can fit them to my own system and stack and make them do what I want, which is usually indexing, associations and repetitive work. I also have a model I have post-trained, and it knows how to disagree and say what it thinks when I ask it to compare data, instead of saying yes to everything. I do not write text with AI, I do not ask it questions and I do not generate images. A diagram at most. Trusting what a model answers, even the best one, depends on the variables: with few of them it will probably get it right, but as soon as there are different variables that also interact outside the objective, that is too much for today's AI, and it still answers with a false sense of certainty, even if you keep it in a sandbox and it queries your sources directly or through vector databases.

That is the danger. A year ago the hallucinations stood out and looked very bad. Now it is worse, because they have been camouflaged: there are fewer of them, but they are still there, and now they carry the confidence of an informed fool when something does not apply or is out of place. The model has to give an answer, and if it finds nothing better it gives you the one that goes off at a tangent. So whoever consults it without backing or cross-checking ends up causing disasters, or building applications they believe are right because they never looked at the code or the assumptions.

For there to be rich people there probably have to be poor ones. Field work pays little for what it demands and for the responsibility you sign off on, and anyone saying otherwise is not out on the farm. Better to know it before getting into it than to find out afterwards. I always say this is like being a vet: it is a vocation, working hard with barely enough, and often not even that is fair. Anyone without it works that out quickly and looks for something quieter.

Even so it is worth it if you like this: you decide about living things, you see the result of what you recommend and no two years are alike. And part of why it pays badly is down to the system. When that advice is driven by the official gazette, the need stops being about productivity for profitability and becomes an obligation that may well hurt my profitability: I have technicians because the law says so, not because they are useful to me. It happens on the commercial side too, where with so much competition everyone offers advice, which is often more of a conversation every two or three weeks. You find growers with five technicians and not one of them doing the job properly.

More than people think, and that bothers me. Fertilisers are not all the same and neither are the products, and with microbiological inputs, which is where things are heading, proving real efficacy is harder still. The difference is not only the cost, it is the composition. Part of the problem is that companies do not know how to show the real value of their products or explain why they stand out: you find firms with better products, which cost them more, earning less than others when it should be the other way round. Real technical selling has almost disappeared, and it is technical not because of the seller's degree but because there is follow-up behind it. There is less of that every year and more cases of not even bothering to follow the product. And if you do not give it value, the technician or the grower ends up caring as little about the product as about the sales technician.

Yes. The site is focused on Spain and on intensive production, but agriculture as a base is the same everywhere and fertigation is what it is. The difference is that some decisions are taken on other technical criteria, and sometimes beyond the technical ones, depending on the social and economic situation of each place. I have added functions to the tools, such as automatic alkalinity control, that we would not use in most of Spain.

Sometimes it makes little difference and other times far more than it should, especially in early detection. With fungal diseases it is very common to use broad-spectrum fungicides, but even those fungicides often have very different efficacy depending on the crop and the fungus. There are many cases of carrying on applying fungicides that harden the plant, add active substances or forcibly stop the electron transporters, with immediate consequences, and on top of that you apply several times and the problem barely goes away because the fungus was not identified correctly, when with the right product it could have been done first time. This matters even more when entomopathogenic fungi are used in biocontrol, because fungicides are very persistent.

A book or a lecture will tell you no, and they have their reasons, but in reality it depends. The mind works in complex ways: we recognise patterns unconsciously. When you go into a crop and you know its context, what might be «this one is short of x» can be clearly reinforced by that context of what has been done, past experience, cases you have seen and so on. All of that arrives at once, and it is also what helps you get it right more often with experience. But it still has two weak points: first, you depend on experience and knowledge; second, missing variables or not considering one can take you down the wrong path. That is why everyone says to analyse. On that point, nobody analyses everything all the time: sometimes only when it is needed, other times continuously, which for monitoring is necessary, because there is a lot of profitability at stake and in the end the cost is nothing. But it has happened to me, and to everyone: being almost certain it is one thing and then, after analysing the plant, the nutrient solution and so on, finding it was something else.

That is the risk you run with eyeballing and experience: if you are missing something important from the context, that is when the errors that cost money come.

I get asked this often, that I am a rare case because there are not many people who combine the field and advisory work, above all, with the digital and technological side in this way. But, as the saying goes, they do exist: plenty of people simply prefer to devote themselves to one or the other, or to neither. As with everything, there is no mystery: everything I learn is because it has a bearing on my work, it is time and will: learning, reading, talking with colleagues, trying things in the field, weighing opinions. You always have to look at things critically, my own included.

Yes, for years now. Nothing special, just to build myself tools. AI now lets me make better interfaces, because I never wanted to spend time on that, since what I program is of direct use for technical work.

If you really want to work in digital agriculture, start with the agriculture first and then the digital. Doing it the other way round is what has led us to where we are: anyone can say they do digital agriculture, but few can say they do agriculture AND digital. That is the idea.

Yes, I have put together a page of books, manuals and databases grouped by topic.

On the one hand it is understandable from the job market's point of view, but on the other it does not hold up: it has diminishing returns, and very quickly. The kind of specialisation that goes «I want someone with 6 years' experience in cherry tomato, variety whatever». It is probably not a good idea for you or for the company to go that way, because they lose the flexibility of having workers who adapt to things and are more experienced.

We are in a time of changing plans, restructuring, acquisitions and working out where workers fit with the products or services on offer. Flexibility and broad knowledge are useful and almost a matter of survival to avoid nasty surprises. I think wasting a profession like engineering on such specific things is bad for everyone: we create professionals who then do not cover companies' new needs and cannot find a place elsewhere.