Method for estimating a vertical profile of relative atmospheric humidity

Abstract

METHOD FOR ESTIMATING A VERTICAL PROFILE OF ATMOSPHERIC RELATIVE HUMIDITY Method (500) for estimating a vertical relative humidity profile (11a-b,30), said method using an input dataset (200), comprising at a minimum an estimated vertical relative humidity profile VPRHe(t) (10a-b, 20), an estimate of cloud cover CFe(t) (21) and an observation of cloud cover CFo(t) (22), said method being implemented by means of a computer device (800), said method comprising the following consecutive steps: (520) identification of local maxima RHmaxi(t) (15a-b) of the vertical relative humidity profile VPRHe(t) (10a-b,20); (530) verification of variability in the observed cloud fraction ∆CFo(k), over a time interval ; and (540) identification of the type of correction that is applied to the VPRHe(t) profile (10a-b,20). Figure from the abstract: figure 4

Description

Title of the invention: Method for estimating a vertical profile of the relative humidity of the atmosphere Technical field

[0001] The present invention belongs to the technical field of weather forecasting. The invention relates more particularly to a method for estimating a vertical profile of the relative humidity of the atmosphere, in order to predict the presence or absence of clouds in a study region and in the short term.

[0002] The invention finds a direct application in the meteorology sector, in particular using the vertical profile of relative humidity in a numerical weather forecast model. State of the art

[0003] Weather models have undergone significant evolution since their development in the 1920s, with the addition of more complex mathematical equations and the use of numerical models to simulate the atmosphere more accurately.

[0004] Since the middle of the 20th century, numerical weather prediction (NWP) models have been developed, allowing longer-term forecasts.

[0005] For example, the GFS (Global Forecast System) model is one of the most widely used NWP models worldwide. The GFS model is developed by the NCEP (National Centers for Environmental Prediction) and can, among other things, provide weather forecasts for a period of up to two weeks.

[0006] Météo-France, the official French meteorological and climatological service, currently uses the NWP models ARPEGE (Action de Recherche Petite Echelle Grande Echelle) and AROME (Applications de la Recherche à l'Opérationnel à Méso-Echelle), which provide high-resolution forecasts over Europe as well as on a global scale for the ARPEGE model, or high-resolution forecasts over a specific geographical area for the AROME model.

[0007] NWP models use high-quality observational data to provide forecasts of the highest possible accuracy, up to several days in advance.

[0008] However, despite their sophistication, current weather models still have limitations and drawbacks.

[0009] The main limitations are initialization errors of observational data, limited spatial and temporal resolution, uncertainties in data and parameters, and the complexity of modeling the atmosphere itself. These limitations can lead to errors in weather forecasts, especially for small-scale weather phenomena and longer-range forecasts.

[0010] Initialization errors in observational data can have significant effects on forecasts. Observational data do not always cover areas of interest by not being available at certain geographic coordinates or altitudes. Also, some data may be inaccurate. In addition, the spatial resolution of models may not be sufficient to predict small-scale weather phenomena, such as thunderstorms or local storms.

[0011] The temporal resolution of weather models is also limited. Current models can forecast several days in advance, but accuracy decreases as the forecast period lengthens. Uncertainties in data and parameters can also lead to uncertainties in forecasts.

[0012] Finally, more accurate modeling of the atmosphere represents a significant challenge for the international meteorological community, as the atmosphere is an extremely complex system, with non-linear interactions between many different processes. Current weather models can simulate many aspects of this complexity, but there is still much to learn about how the atmosphere works.

[0013] Thus, a better evaluation of the input data of PNT models represents an axis of improvement in the modeling of the dynamic behavior of the atmosphere.

[0014] Document WO2016046309Al describes an assembly and a detection method for identifying and tracking clouds in an area of ​​the sky. One of the steps of the method consists in particular in deducing a vertical distribution of water vapor in a given location, and this from measurements taken at ground level, as well as from previously modeled atmospheric profiles. On the other hand, document WO2016046309Al does not describe the steps for correcting the vertical profile of relative humidity, nor does it provide them.

[0015] Document US2020257018A1 describes systems and methods for measuring environmental parameters, including air temperature and humidity. In one embodiment, a measurement system comprises a mast that is equipped with a plurality of sensors, positioned at different heights along the length of the mast. Thus, a local vertical profile of parameters relating to the characterization of the air is obtained. On the other hand, the system described in document US2020257018A1 is characterized by a maximum study height from the ground that is of the order of 3 m, which is not sufficient for PNT models. Furthermore, other documents describe measurement masts that can reach several hundred meters in height, without that these devices are sufficient to estimate a vertical profile of relative humidity for the entire tropopause: the tropopause corresponding to the layer of the atmosphere in which meteorological phenomena are created and studied.

[0016] Document GB2574154A describes a method for obtaining vertical profiles of several atmospheric parameters in a synchronized manner. Thus, document GB2574154A describes the use of different measuring devices based on various principles capable of measuring several parameters at the same time, said devices being fixed on a cable of an aerostat at a specified distance interval. Furthermore, each measuring device is capable of transmitting in real time the collected data to a device for synchronous acquisition of profiles, measured in real time, of several atmospheric parameters, said device being located on the ground. Although the measurements made can be obtained up to high altitude, the implementation of the device is not easy, can be restrictive, in particular depending on extreme weather conditions.

[0017] It is therefore necessary to continue the development of easy-to-implement solutions, which can be used continuously, day and night without constraints, and which make it possible to obtain precision as well as a rapid evaluation of the input data of the PNT models, in particular the vertical profile of relative humidity.

[0018] The vertical profile of relative humidity can be estimated by determining the presence of clouds, their heights as well as their thicknesses.

[0019] Document CN112085751A describes a method for estimating the height of the cloud layer based on cloud image shadow matching algorithm, which includes the following steps: - synchronized shooting of a cloud image and a ground shadow image and preprocessing; - selecting a cloud cluster matching point and a shadow matching point by means of the cloud cluster shadow matching algorithm; and - resolution of the height of the cloud layer based on the actual spatial relationship of the correspondence points.

[0020] Document WO2016030951Al describes an automated sky observation device comprising, among other things, a light emission / reception unit, a unit for correcting the measured signal, a cloud layer estimation unit, a cloud base altitude estimation unit and a cloud cover estimation unit.

[0021] On the other hand, it would seem that the estimation method described in document CN112085751A, and the device described in document WO2016030951A1, are essentially based only on a measured quantity to evaluate the presence, altitude, and cloud thickness in an area of ​​interest. However, this approach may not be sufficiently accurate for applications in meteorology in particular, where other input data are needed to refine this estimate.

[0022] Thus, it is difficult to determine the height where clouds are located over an area of ​​interest, and it is therefore necessary to develop solutions improving these estimates. Presentation of the invention

[0023] The present invention aims to overcome all or part of the drawbacks presented above, and then proposes an innovative solution making it possible in particular to correct the altitude as well as the vertical spread of clouds present in a vertical atmospheric profile at a geographical coordinate.

[0024] To this end, the present invention aims at a method for estimating a vertical profile of relative humidity, said method using an input data set, comprising at least an estimated vertical profile of relative humidity VPRHe(t), an estimation of the cloud cover CFe(t) and an observation of the cloud cover CFo(t), said method being implemented by means of a computer device.

[0025] This estimation method is remarkable in that it comprises the following consecutive steps: - identification of local maxima RHmaxi(t) of the vertical relative humidity profile VPRHe (t); - verification of a variability of the observed cloud fraction ACFo(k), over a time interval; and - identification of the type of correction which is applied to the profile VPRHe(t).

[0026] According to a particular characteristic, the estimation method further comprises, and before the step of identifying local maxima RHmaxi(t), a step of verifying a difference ACFe o(t) between the estimate of the cloud cover CFe(t) and the observed cloud cover CFo(t).

[0027] According to another particular characteristic, the estimation method further comprises, and following the step of identifying the type of correction, a step of adding a cloud and / or a step of removing a cloud, at the altitude of the local maximum RHmax^t).

[0028] According to another particular characteristic, the estimation method further comprises a step of preserving the VPRHe(t) profile.

[0029] According to another particular characteristic, the estimation method further comprises a step of generating a corrected vertical relative humidity profile VPRHc(t).

[0030] Advantageously, the step of identifying local maxima RHmaxft) of the vertical relative humidity profile VPRHe (t) uses an observation of the height of the cloud base CBHo (t).

[0031] In another advantageous manner, the step of identifying local maxima RHmaxi(t) of the vertical profile of relative humidity VPRHe uses an observation of the height of the top of the clouds CTHo(t) of the input data set.

[0032] Advantageously, the step of identifying local maxima RHmaxft) of the vertical relative humidity profile VPRHe uses an observation of the type of clouds in the input data set.

[0033] According to another particular characteristic, the step of identifying local maxima RHmaXi(t) of the vertical relative humidity profile VPRHe uses a plurality of information, such as a vertical spread of the clouds and / or an optical thickness of the clouds.

[0034] Another object of the invention is a computer program product remarkable in that it comprises a set of program code instructions which, when executed by a processor, implement an estimation method as described above.

[0035] Finally, the present invention also relates to a computer device, implementing an estimation method as described previously, characterized in that said device comprises a memory, a data processor, a means for transferring incoming and outgoing information.

[0036] The fundamental concepts of the invention having been set out above in their most elementary form, other details and characteristics will emerge more clearly on reading the description which follows and with reference to the appended drawings. Presentation of the drawings

[0037] The figures are given purely for illustrative purposes for a better understanding of the invention without limiting its scope. The various elements may be represented schematically and are not necessarily on the same scale. Throughout the figures, identical or equivalent elements bear the same numerical reference.

[0038] It is thus illustrated in:

[0039] [Fig. 1]: a graph representing a first vertical profile of the relative humidity and the same profile corrected by an estimation method, according to one embodiment;

[0040] [Fig.2]: a graph representing a second vertical profile of the relative humidity and the same profile corrected by the estimation method, according to one embodiment;

[0041] [Fig.3]: a diagram in the form of a diagram of the input data of the estimation method and the output data;

[0042] [Fig.4]: the main steps of the method for estimating the vertical profile of relative humidity according to one embodiment of the invention;

[0043] [Fig.5]: a diagram in the form of a diagram of a computer device allowing the implementation of the estimation method. Detailed description of embodiments

[0044] It should be noted that certain technical elements well known to those skilled in the art are described herein to avoid any insufficiency or ambiguity in the understanding of the present invention.

[0045] In the embodiment described below, reference is made to a method for estimating a vertical profile of relative humidity, said profile being used as input data in numerical weather prediction (NWP) models.

[0046] This estimation method has the particular advantage of updating the vertical profile of relative humidity in near real time and continuously, thus making it possible to obtain more precise input data that is closer to reality, thus improving weather forecasts.

[0047] [Fig.l] and [Fig.2] represent an example of graph 100, each presenting a vertical profile of the relative humidity VPRHe(t) lOa-b estimated at a time t and for a geographic coordinate.

[0048] In this embodiment, the graph 100 comprises an abscissa axis 18, which expresses the VPRHe(t) lOa-b profile as a percentage, on a numerical scale ranging from 0% to 100%, as well as an ordinate axis 19 which represents the altitude, above ground level, and which is expressed in kilometers on a numerical scale ranging from 0 km to 30 km.

[0049] In other embodiments, the altitude at which the VPRHe (t) lOa-b profile is estimated may be expressed using other units, in pressure level (unit hPa) or using a number of digital layers of the model.

[0050] [Fig.l] and [Fig.2] each also represent, by means of dashed-dot lines, a corrected vertical relative humidity profile VPRHc(t) lla-b. The profile VPRHc(t) lla-b corresponds to the profile VPRHe(t) lOa-b which has been modified by means of a method 500 for estimating a vertical relative humidity profile, the main steps of which, and according to one embodiment, are represented in [Fig.4].

[0051] One of the main objectives of the estimation method 500 is to correct the profile VPRHe(t) lOa-b, based in particular on different observations of the sky, in order to correct the relative humidity RH around the altitude at which clouds were observed. This is the example illustrated in [Fig.l], where a local maximum of relative humidity RHmax^t) 101a, of the profile VPRHe(t) 10a, of a value close to 85%, is corrected so that in the profile VPRHc(t) 1 la, a local maximum of relative humidity RHmax; (t) 11 la, has a threshold value of 100%, which unequivocally indicates the presence of a cloud at the corresponding attitude, which has been previously confirmed at an altitude 14, by means of a sensor, said altitude being represented in [Fig.l] by a dotted line.

[0052] Another objective of the estimation method 500, illustrated in [Fig.2], is to correct the profile VPRHe(t) lOa-b by reducing the value of a local maximum of relative humidity RHmaxi(t) 101b, of the profile VPRHe(t) 10b, so that in the profile VPRHc(t) 11b, a local maximum of relative humidity RHmax^t) 111b, has a value which does not allow to have a cloud at the corresponding attitude.

[0053] This analysis of the relative humidity RH is carried out between a minimum study altitude Altmin 13 and a maximum study altitude Altmax 12.

[0054] [Fig.3] represents a diagram detailing an input data set 200, according to an embodiment of the invention, to be used by the estimation method 500 to generate an output data set 300, which in this embodiment is a vertical profile of relative humidity VPRHc(t) 30, at a time t, as well as at a geographic coordinate.

[0055] In this particular embodiment of the invention, the input data set comprises: - an estimated vertical profile of relative humidity VPRHe(t) 20, expressed as a percentage, and which may come for example from a PNT model, without this representing a limit to the invention; - an estimate of cloud cover CFe(t) 21, expressed as a percentage; - an observation of the cloud cover CFo(t) 22; - an observation of the height of the cloud base CBHo(t) 23; - an observation of the height of the cloud tops CTHo(t) 24; - an observation of cloud type 25; - as well as a plurality of information 26 making it possible to characterize the physical parameters of the clouds, such as the vertical spread of the clouds, the optical thickness.

[0056] It is recalled here that the input data set 200 is obtained for a single geographic coordinate in order to carry out the estimation of the vertical profile of relative humidity, and this at a time t.

[0057] The estimation of the cloud cover CFe(t) 21 is for example deduced from a PNT model, without this representing a limit to the invention.

[0058] In addition, the observation of the cloud cover CFo(t) 22 is, in a particular embodiment, obtained from a sensor developed by the company Reuniwatt, the Sky InSight™ infrared sky imager. Other methods known to those skilled in the art can also provide observation of cloud cover CFo(t) 22, such as images obtained using cameras or satellites (geostationary and / or polar-orbital).

[0059] In a particular embodiment of the invention, the profile VPRHe(t) 20, the estimation of the cloud cover CFe(t) 21 and the observation of the cloud cover CFo(t) 22 are input data which are mandatory for executing the estimation method 500.

[0060] In another particular embodiment of the invention, the observation of the height of the base of the clouds CBHo(t) 23, for example by means of ceilometers, the observation of the height of the top of the clouds CTHo(t) 24, for example by means of satellites, are data which are optional for executing the estimation method 500.

[0061] Similarly, in another particular embodiment of the invention, the observation of the type of clouds 25 is obtained by means of readings carried out by an operator, or images originating from cameras and / or satellites analyzed by means of artificial intelligence algorithms, without this being a limit of the invention, the person skilled in the art being able to use other known methods.

[0062] Finally, the input data set 200 can be supplemented by a plurality of information 26 coming in particular from ground weather stations 26a, weather balloons 26b as well as satellites 26c, information that a person skilled in the art is able to supplement.

[0063] [Fig.4] represents the main steps of a method of estimating 500 the profile vertical relative humidity VPRHc(t) 30, according to one embodiment, and such that said profile is represented in figures 1 and 2.

[0064] The method 500 mainly comprises: - a step 510 of verifying a difference ACFeo(t) between the estimated cloud cover CFe(t) 21 and the observed cloud cover CFo(t) 22; - a step 520 of identification of the local maximums RHmax^t) lOla-b in the profiles VPRHe(t) lOa-b; - a step 530 of verifying a variability of the observed cloud fraction ACFo(k), over a time interval such that a time k is between a time tx and a time t+y; - a step 540 of identifying the type of correction which is applied to the profile VPRHe(t) lOa-b; - a step 550 of adding a cloud at the altitude of the local maximum RHmaXi(t) lOla-b, or a step 560 of removing a cloud at the altitude of the local maximum RHmaxi(t) lOla-b; - a step 570 of preserving the profile VPRHe(t) lOa-b or a step 580 of generating the profile VPRHc(t) 1 la-b.

[0065] More precisely, step 510 of verifying the difference ACFeo(t) between the estimated cloud cover CFe(t) 21 and the observed cloud cover CFo(t) 22 consists of comparing the value of ACFeo(t) with constants ACFmin and ACFmax.

[0066] Thus, if the difference ACFe „(0 is less than the constant ACFmin, or greater than the constant ACFmax, the profile VPRHe(t) lOa-b is not corrected, and the estimation method 500 of the vertical profile of relative humidity 30 stops while retaining the profile VPRHe(t) 20 used as input to said method.

[0067] On the other hand, and if the difference ACFe „(0 is between the constant values ​​ACFmin and ACFmax, the estimation method 500 continues.

[0068] Step 520 consists of identifying the local maximum relative humidity RHmax; (t) lOla-b which are between the minimum analysis altitude Altmin 13 and the maximum analysis altitude Altmax 12. For each local maximum of relative humidity RHmaxi(t) lOla-b, an altitude AltRHmaxi(t) is associated.

[0069] Then, the values ​​of the local maximum relative humidity RHmax^t) lOla-b are compared to a minimum threshold of relative humidity RHmin. If the value of the local maximum relative humidity RHmaxi(t) 10la-b is lower than RHmin, the method 500 for estimating the vertical profile of relative humidity 30 stops while retaining the profile VPRHe(t) 20 used as input to said method.

[0070] On the other hand, if the value of the local maximum relative humidity RHmax^t) lOla-b is greater than RHmin, the estimation method 500 continues with a view to potentially applying a correction to the input profile VPRHe(t) 20.

[0071] According to a particular embodiment of the invention, if: - observation of the height of the cloud base CBHo(t) 23; - observation of the height of the cloud tops CTHo(t); 24 and / or - observation of cloud type 25;

[0072] are available in the input data set 200, then the differences between the values ​​of the altitudes AltRHmax^t) (corresponding as a reminder to local maximums of relative humidity RHmax^t) 10la-b) and the heights of the base and the top of the clouds, respectively CBHo(t) 23 and CTHo(t) 24, are calculated.

[0073] If this difference is greater than a maximum altitude step AAltmax, and this for all the data pairs {RHmaxft); AltRHmax^t)}, the method estimation 500 of the vertical profile of relative humidity 30 stops while keeping the profile VPRHe(t) 20 used at the input of said process.

[0074] Otherwise, steps 530, 540, 550 and 560 are applied to the set of data pairs {RHmaxi(t); AltRHmaxi(t)} in order to refine the vertical relative humidity profile 30 at the output of the estimation method 500.

[0075] According to a particular embodiment of the invention where the observation of the height of the cloud base CBHo(t) 23, the observation of the height of the cloud top CTHo(t) 24 as well as the observation of the type of clouds 25 are not available in the input data set 200, no comparison is then made between the AltRHmaXi(t) corresponding to local maximums of relative humidity RHmax^t) lOla-b, and for which the presence of a cloud is a priori assumed.

[0076] Step 530 consists of verifying the variability of the observed cloud fraction ACFO (k) in order to avoid having to modify the input profile if the weather conditions are uncertain (in other words varying too quickly). This verification is carried out over a time interval {tx; t+y], which in this particular embodiment is 45 min long.

[0077] The variability of the observed cloud fraction ACFo(k) is evaluated using the following formula, without this presenting a limit to the invention:

[0078] [Math.l] ACFdk) = max CFJæI- min CfJaj \ / tx <k<t+y \ / t-x<k<t+y \ J

[0079] Thus, the difference between the maximum of the observation of the cloud cover CFo(k) 22 and its minimum is evaluated over the entire time interval {tx; t+y]. Once the variability of the observed cloud fraction ACFo(k) has been calculated, it is compared to a maximum variability threshold ACFspreadmax, and if ACFo(k) is greater than said threshold, the estimation method 500 of the vertical profile of relative humidity 30 stops, retaining the profile VPRHe(t) 20 used as input to said method. Indeed, it is then chosen not to modify the vertical profile of relative humidity 30 because the cloud fraction varies too quickly over the observed time scale. This would therefore not provide more precision to the PNT model and would risk causing poor initialization of the PNT model during the next iteration.

[0080] Step 540 consists of identifying the type of correction which is applied to the VPRHe(t) lOa-b profile.

[0081] According to a particular embodiment of the invention, this identification is carried out by first comparing the deviation ACFeo(t) with the minimum difference threshold ACFmin, as well as the values ​​of the local maximum relative humidity RHmaxi(t) lOla-b with a minimum reference relative humidity threshold RHref.

[0082] Thus, if the difference ACFeo(t) is greater than the minimum difference threshold ACFmin, the vertical profile of relative humidity is modified locally, by increasing the relative humidity around the local maximum relative humidity RHmaxi(t) lOla-b concerned. This is step 550 of adding a cloud.

[0083] In the case where the deviation ACFe „(0 is less than the minimum difference threshold ACFmin, and if the values ​​of the local maximums of relative humidity RHmaxi(t) lOla-b are greater than the minimum threshold of reference relative humidity RHref, the vertical profile of relative humidity is modified locally, by reducing the relative humidity around the local maximums of relative humidity RHmaxft) lOla-b concerned, and according to a reduction of the relative humidity ARH. This is the step 560 of removing a cloud.

[0084] For the other cases, the method of estimating 500 the vertical relative humidity profile 30 stops by retaining the profile VPRHe(t) 20 used as input to said method, in its entirety or, only around the altitudes AltRHmaxi(t) associated with the local maximums of relative humidity RHmaxft) lOla-b which do not satisfy the criteria presented.

[0085] It is specified that the method, described here, for identifying an addition or a deletion of a cloud in the vertical profile of relative humidity, does not represent a limiting aspect of the invention, and that other criteria and calculations can in particular be used.

[0086] As mentioned previously, step 550 consists of adding a cloud by modifying the profile VPRHe(t) 10a at the altitude AltRHmaxi(t) of the local maximum relative humidity RHmaxft) 101a, so that said local maximum has a threshold value of 100%. This correction can be carried out by means of various mathematical functions, such as a Gaussian or Lorentzian function. This correction is represented in [Fig.l], at the local maximum relative humidity RHmaxft) 11 la of the profile VPRHc(t) 11a.

[0087] Step 560 consists of removing a cloud in the vertical relative humidity profile. As previously, this correction can be carried out by means of various continuous mathematical functions, or by means of reducing the relative humidity ARH, defined by relative humidity range RH. This correction is shown in [Fig.2], at the level of the local maximum relative humidity RHmaxi(t) 111b of the profile VPRHc(t) 11b.

[0088] Finally, steps 570 and 580 consist respectively of either: - retaining the vertical relative humidity profile 20 from the input data set 200, which then becomes the vertical relative humidity profile 30 of the output data set 300; - generate a vertical profile of relative humidity 30, of which local maximum relative humidity RHmaxft) have been corrected (either increased to 100%, or decreased).

[0089] By way of example, and without this presenting a limit to the invention, table 1 below summarizes the values ​​used by the estimation method 500.

[0090] [Tables 1] Parameter Value Minimum difference threshold ACFmin -0.5 Minimum difference threshold ACFmax 0.5 Minimum analysis altitude Altmin 1000 m Maximum analysis altitude Altmax 12000 m Minimum relative humidity threshold RHmin 60% Maximum altitude step AAltmax 3000 m Upstream time step x 30 min Downstream time step y 15 min ACFspreadmax 0.65 Minimum reference relative humidity threshold RHref 85% Relative humidity reduction ARH if 75% < RH < 85%, AHR = 5% if 85% < RH < 95%, AHR = 10% if 95% < RH, AHR = 20%

[0091] For a particular embodiment, the method 500 is implemented by a device 800 as described below.

[0092] [Fig.5] represents a functional diagram of the computing device 800 for implementing one or more embodiments of the invention.

[0093] The computing device 800 may be a device such as a server, or a computing unit, which does not require, in the preferred embodiment, the intervention of an operator for the implementation or execution of the steps of the estimation method 500.

[0094] In a particular embodiment, the computing device 800 comprises a memory 81, for storing program instructions as well as data, a data processor 82 capable of executing at least one step of the estimation method 500, and a means for transferring incoming and outgoing information 83, such as a wired or wireless network interface.

[0095] This particular embodiment is given by way of example, and does not represent a limit to the present invention. Indeed, to implement the estimation method 500, a person skilled in the art is able to use computer devices having functionalities that are all or partly identical to those described previously.

Claims

Claims

1. Method for estimating (500) a vertical relative humidity profile (1 la-b,30), said method using an input data set (200) obtained at a time t and for a geographical coordinate, comprising at least an estimated vertical relative humidity profile VPRHe(t) (10a-b,20), an estimation of the cloud cover CFe(t) (21) and an observation of the cloud cover CFo(t) (22), said method being implemented by means of a computer device (800), said method being characterized in that it comprises the consecutive steps: - (520) of identifying local maxima RHmaxft) (15a-b) of the vertical relative humidity profile VPRHe(t) (10a-b,20);- (530) of verification of a variability of the observed cloud fraction ACFo(k), the variability of the cloud fraction being the difference between the maximum of the observation of the cloud cover CFo(k) 22, and its minimum over a time interval such that a time k is between a minimum time tx and a maximum time t+y where x and y are delays; and - (540) of identification of the type of correction which is applied to the profile VPRHe(t) (10a-b,20), the type of correction consisting of decreasing or increasing the relative humidity around local maximums of relative humidity RHmaxi(t) lOla-b, in the case where a correction is relevant.;

2. Method (500) according to claim 1, further comprising, and before the step (520) of identifying local maxima RHmaxi(t) (15a-b), a step (510) of verifying a difference ACFeo(t) between the estimate of the cloud cover CFe(t) (21) and the observed cloud cover CFo(t) (22);

3. Method (500) according to claim 1 or 2, further comprising, and following the step (540) of identifying the type of correction, a step (550) of adding a cloud and / or a step (560) of removing a cloud, at the altitude of the local maximum RHmaxi(t) (15a-b).

4. Method (500) according to claim 1 or 2, further comprising a step (570) of preserving the VPRHe(t) profile (10a-b,20).

5. Method (500) according to claim 1, 2 or 3, further comprising a step (580) of generating a corrected vertical relative humidity profile VPRHc(t) (lla-b,30).

6. Method (500) according to any one of the preceding claims, wherein the step (520) of identifying local maxima RHmaXi(t) (15a-b) of the vertical profile of relative humidity VPRHe (t) (10a-b,20) uses an observation of the height of the base of the clouds CBHo(t) (23) of the input data set (200).

7. Method (500) according to any one of the preceding claims, wherein the step (520) of identifying local maxima RHmaXi(t) (15a-b) of the vertical profile of relative humidity VPRHe (t) (10a-b,20) uses an observation of the height of the top of the clouds CTHo(t) (24) of the input data set (200).

8. Method (500) according to any one of the preceding claims, in which the step (520) of identifying local maxima RHmaxi(t) (15a-b) of the vertical profile of relative humidity VPRHe (t) (10a-b,20) uses an observation (25) of the type of clouds of the input data set (200).

9. Method (500) according to any one of the preceding claims, in which the step (520) of identifying local maxima RHmaXi(t) (15a-b) of the vertical profile of relative humidity VPRHe (t) (10a-b,20) uses a plurality of information (26) characterizing physical parameters of clouds, such as a vertical spreading of the clouds and / or an optical thickness of the clouds.

10. Computer program product characterized in that it comprises a set of program code instructions which, when executed by a processor (82) of a computing device (800), implement an estimation method (500) according to one of claims 1 to 9.

11. Computer device (800), implementing an estimation method (500), according to any one of claims 1 to 9, characterized in that said device comprises, a memory (81), a data processor (82), a means for transferring incoming and outgoing information (83).

Images