Is the weather forecast accurate for 40 days?

Accurate, the principle of modern weather forecast and climate forecast is the result of numerical model. To put it bluntly, the numerical model is a set of equations describing the atmospheric motion, and the numerical results are obtained by computer calculation. But these results can't directly reflect the weather phenomenon. On the contrary, it reflects various variables of the atmosphere (pressure, temperature, wind speed, precipitation, etc. Forecasters or researchers combine these variables to predict the possible weather. There is a problem that needs to be explained. There are two ways to drive numerical model values. One is to enter the initial variables and then see the changes in the next few days. For example, enter various meteorological variables (those mentioned above) today and see what these variables will be tomorrow. This is the basic principle of daily weather forecast. The accuracy of this forecast in 7- 10 days is basically acceptable. If it exceeds 10 days, with the current technical conditions, the error will be very large, almost the same as blindness. Then someone asked, climate prediction, a prediction is months and years, isn't that even more unreliable? Actually, it's about another driving mode, that is, inputting boundary conditions, which is a bit silly? No, boundary conditions basically refer to topography, seawater temperature, solar radiation and other variables. Also called external force. Under these external forcing conditions, after a certain period of time, the climate model can reach the corresponding balance. Researchers can predict the future climate through the equilibrium state. Of course, this kind of climate forecast is not the daily weather condition, but the climate condition on a longer time scale. For example, will it be cold or warm in the northeast next spring? Is it drought or waterlogging? As for which day there will be high/low temperature and which day there will be heavy rain, we can't report it. But this qualitative information is already very useful. For example, in autumn, if it is known in advance whether winter is cold or warm, down jacket manufacturers can decide the stocking quantity (the example given by the teacher in class is a bit low, but it is still quite suitable). First of all, the accuracy of daily weather forecast decreases with the increase of forecast time. In other words, the longer the forecast time, the less accurate the month. On the contrary, the longer the input forced climate prediction model is, the more accurate it is (because it takes time for the model to reach equilibrium, and the time is short, which may be unbalanced). There is an awkward point in the middle, that is, the forecast time is within the range of 10-30 days, and it is not allowed to be day by day, and neither is the climate forecast. This is called long-term prediction, which is a difficult point in scientific research. So, it's almost enough to watch it for seven days every day.