SIDC/SILSO Latest sunspot number forecast

The SIDC/SILSO (Sunspot Index and Long-term Solar Observations) produces 12 months ahead predictions of the monthly smoothed sunspot number using three different methods. In addition, for each of the methods, there is also a Kalman filter optimised version available.


Issue time of the data: 01 June 2021

 Methods
dateML2SC2CM2ML2KFSC2KFCM2KF
2020-1215.8±2.415.4±11.116.1±3.215.8±0.016.0±0.016.6±0.0
2021-0118.2±4.817.9±13.919.6±3.918.2±0.017.4±0.018.8±0.0
2021-0220.7±7.618.6±16.722.8±4.620.7±0.016.3±0.019.1±0.0
2021-0324.1±9.721.0±19.427.1±5.419.3±2.718.9±2.620.1±2.8
2021-0428.2±13.122.6±22.232.2±6.422.4±3.320.5±3.023.6±3.5
2021-0532.5±18.124.5±25.036.7±7.325.4±4.122.2±3.626.4±4.2
2021-0636.7±22.626.5±27.840.9±8.228.7±5.124.0±4.429.5±5.3
2021-0740.8±26.928.4±30.645.6±9.131.9±6.225.7±5.332.9±6.4
2021-0844.8±30.730.7±33.351.0±10.235.1±7.427.8±6.236.8±7.7
2021-0949.2±34.632.8±36.157.2±11.438.5±8.529.7±7.041.2±9.1
2021-1054.3±39.335.1±38.963.5±12.742.5±9.931.8±7.945.8±10.5
2021-1158.5±43.837.2±41.770.1±14.045.8±11.133.7±8.850.5±12.0
2021-1262.4±48.039.7±44.476.3±15.348.9±12.236.0±9.855.0±13.5
2022-0167.4±53.141.8±47.281.1±16.252.8±13.637.9±10.658.4±14.7
2022-0271.9±59.243.3±50.085.0±17.056.3±14.939.2±11.461.3±15.8
2022-0375.4±64.246.4±52.888.3±17.759.0±16.042.0±12.563.6±16.8
2022-0479.6±66.049.9±55.692.2±18.462.3±17.245.2±13.866.4±18.0
2022-0584.1±66.654.2±58.395.8±19.265.8±18.549.1±15.369.0±19.0

This web page forms part of the ESA Space Situational Awareness Programme's network of space weather service development activities, and is supported under ESA contract number 4000113186/15/D/MRP.

For further product-related information or enquiries contact helpdesk. E-mail: helpdesk.swe@ssa.esa.int

All publications and presentations using data obtained from this site should acknowledge Solar Influences Data analysis Centre (SIDC) at the Royal Observatory of Belgium (ROB) and The ESA Space Situational Awareness Programme.

For further information about space weather in the ESA Space Situational Awareness Programme see: www.esa.int/spaceweather

Access the SSA-SWE portal here: swe.ssa.esa.int

The data contained in the present page are Intellectual Property of the Royal Observatory of Belgium (ROB) and the contributing observing stations and individual observers. Use of the data is restricted by the ROB data policy.

component = latest

Available at:

https://ssa.sidc.be/prod/API/index.php?component=latest&pc=S108&psc=b
(Base url, requires additional paramaters)

Returns sunspotnumber forecast in png or json.

https://ssa.sidc.be/prod/API/index.php?format=png&component=latest&pc=S108&psc=b

https://ssa.sidc.be/prod/API/index.php?format=png&method=SC2KF&component=latest&pc=S108&psc=b

https://ssa.sidc.be/prod/API/index.php?format=json&component=latest&pc=S108&psc=b

Format

JSON example
{
"SunspotNumber_MonthlyForecast": [{,
"data": {
"NominalLeadTime": [number: -5 / 12]
"Uncertainty": [number]
"Value": [number]
"Version": [text v1 or v2]
"Time": [text: yyyyMM]
"Method": [text]
"IssueTime": [datetime]
},
"event_type": "SunspotNumber_MonthlyForecast",
"name": [text],
"id": [number],
"valid_from": [datetime]
}, [REPEAT]
]}

component = archive

Available at:

https://ssa.sidc.be/prod/API/index.php?component=archive&pc=S108&psc=b
(Base url, requires additional paramaters)

Returns a JSON containing the sunspot number monthly forecast data.
at least one of the following parameters is required:

https://ssa.sidc.be/prod/API/index.php?component=archive&pc=S108&psc=b&dts_start=2017-03-10T00:00:00Z&dts_end=2017-04-12T00:00:00Z

https://ssa.sidc.be/prod/API/index.php?component=archive&pc=S108&psc=b&dts_start=2017-03&dts_end=2017-04

https://ssa.sidc.be/prod/API/index.php?component=archive&pc=S108&psc=b&dts_start=2017-03&dts_end=2017-04&nominalleadtime=4

https://ssa.sidc.be/prod/API/index.php?component=archive&pc=S108&psc=b&dts_start=2017-03&dts_end=2017-04&method=CM2

https://ssa.sidc.be/prod/API/index.php?component=archive&pc=S108&psc=b&dts_start=2017-03&dts_end=2017-04&nominalleadtime=4&method=CM2

Format

JSON example
{
"SunspotNumber_MonthlyForecast": [{,
"data": {
"NominalLeadTime": [number: -5 / 12]
"Uncertainty": [number]
"Value": [number]
"Version": [text v1 or v2]
"Time": [text: yyyyMM]
"Method": [text]
"IssueTime": [datetime]
},
"event_type": "SunspotNumber_MonthlyForecast",
"name": [text],
"id": [number],
"valid_from": [datetime]
}, [REPEAT]
]}


SIDC/SILSO Sunspot number forecast

Introduction

Solar activity is governed by the emergence of powerful magnetic fields at the surface of the Sun, marked by sunspots, and it varies cyclically with an average period of about 11 years. The global solar activity level is characterized by the daily number of sunspots, which is available over the past 300 years, and thus includes more than 25 solar cycles. The long-term activity has impacts on planetary environment throughout the solar system, including the Earth’s environment and climate. Regarding human activities, long-term predictions of solar activity are particularly important for space mission planning and lifetime (occurrence rates of particle events, atmospheric drag on low-orbit satellites), and also for aviation (cumulated radiation doses for crews) or maintenance of ground infrastructure like the electrical power grid or pipeline networks (rate and intensity of surges of ground-induced currents).

Prediction methods

The SIDC/SILSO (Sunspot Index and Long-term Solar Observations) produces 12 months ahead predictions of the monthly smoothed sunspot number using three different methods, listed here in the chronological order of their introduction, which is also the order of increasing capabilities:

 

1. McNish & Lincoln (ML) method: Forecasts of the monthly mean sunspot number using the McNish and Lincoln method. The base principle was described in McNish A.G. & Lincoln J.V., 1949, (Transactions, American Geophysical Union). These were the standard solar cycle predictions provided until recently by the Space Weather Prediction Center at National Oceanic and Atmospheric Administration (USA). This method consists of least-square fitting a single mean cycle profile to the last 13-month smoothed sunspot numbers. This mean cycle is obtained by a simple averaging of all cycles between 1849 and 1975 (cycles 9 to 20), aligned at the time of their preceding minimum. This method can provide a reasonable prediction in the course of a cycle and a fair estimate of the upcoming maximum in the rising phase of the cycle. On the other hand, it is unusable around the times of cycle minima and cannot provide any prediction of the expected length of a cycle.

 

2. Standard Curves (SC) method: Forecasts of the monthly mean sunspot number using the Standard Curve method, which is an extension of the original method by M. Waldmeier, 1968 (Astronomy Mitteilungen Eidg. Sternwarte Zurich, N° 286, 1 - 13.). The method consists in the least square fit and interpolation of a set of standard curves, each curve corresponding to the average shape of solar cycles of a narrow range of maximum value. The fit is done on the observed 13-month smoothed monthly sunspot numbers, using the last 24 available values. This is why the prediction actually starts 5 months before the last elapsed month and runs over 18 months (up to 12 months ahead). Compared to the McNish & Lincoln method, the Standard Curves predictions take into account the changing shape (rising/declining rates and durations) of the solar cycle as a function of the maximum amplitude. This method performs well in the middle of each cycle, but as with all methods based purely on past solar activity, it becomes unreliable at the end of each cycle and during the minima.

 

3. Combined method (CM): Forecasts of the monthly mean sunspot number using the Combined Method. The method (Denkmayr & Cugnon 1997, Solar-Terrestrial Prediction Workshop V, eds. G.Heckman et al., Hiraiso Solar Terrestrial Research Center, Japan) consists of building a mean cycle by averaging all past cycles with an amplitude close the amplitude of the cycle to be predicted, among all past recorded cycles since 1749. The mean cycle is calculated dynamically for each prediction, instead of interpolated in a table of pre-calculated cycles like is done for the Standard Curves method above. The fit is done on the observed 13-month smoothed monthly sunspot numbers. This is why the prediction starts 5 months before the last elapsed month and runs over 18 months (up to 12 months ahead). In order to obtain a reliable prediction around the time of cycle minima (from 1.5 years before up to the next maximum), this method uses the aa geomagnetic index as precursor index for predicting the maximum value of the next cycle. The least square fit then uses a combination of the last observed monthly sunspot number and this aa-based prediction of the next maximum value (until it is actually reached). This method thus keeps a predictive capability when passing the minimum phase and early rising phase of a new cycle.

As, in all three above methods, the fit of the predicted values to the last observed monthly smoothed sunspot numbers includes errors due to the randomness of solar activity from month to month, an optimization of the primary predicted values was added by applying an adaptive Kalman filter (T. Podladchikova, R. Van der Linden, 2012, Solar Physics):

 

4. McNish & Lincoln method with adaptive Kalman filter (MLKF)

 

5. Standard Curves method with adaptive Kalman filter (SCKF)

 

6. Combined method with adaptive Kalman filter (CMKF)

A Kalman-optimized prediction is thus available for the first three predictions. They are not independent products as they are largely matching the input predictions, but the Kalman-optimized output adds slight adjustments that smooth out local artificial jumps and improves the continuity between the last observed sunspot number values and the base predicted values.

The uncertainty ranges provided for the predictions correspond to the standard error, derived from the randomness of solar activity over all cycles in the input data. For the McNish & Lincoln method, they are derived mathematically from the input data, while for the other two methods, they were derived empirically from the statistical comparison of hindcasts of past cycles with the actual observed sunspot numbers.

The sunspot numbers reference time series has been recalibrated, and since July 1, 2015, the original Sunspot numbers (Version 1.0) have been replaced by a new entirely revised data series, Version 2.0 (Clette, F. et al., 2014, Space Science Reviews; Clette, F. et al., 2016, Solar Physics). The predictions were based on the original Sunspot number data before July 1, 2015, and are now based on the revised data series since July 1, 2015 (see table below).

 Prediction methods
Sunspot
Number
Version
MLSCCMMLKFSCKFCMKF
1.0/SC1CM1///
2.0ML2SC2CM2ML2KFSC2KFCM2KF

As the scale of the new sunspot number has significantly changed, by a factor of about 1.5), the new predictions are higher and match the scale of the new Version 2.0 time series. One should note that the sunspot numbers Version 1.0 is outdated and now discontinued, and only version 2.0 should be used as a reference for operational applications.

Provided information

This product provides access to the six prediction types for 1 to 12 months ahead predictions of the monthly smoothed sunspot number as explained above (see figures below).

Users can search in the archive, by selecting first the version of the sunspot numbers series, and then the methods (one or more), and the lead time. It will provide a table with the corresponding sunspot number prediction values and the uncertainty.

The version refers to which sunspot number time series the method is making a forecast for: the original (version 1.o) or the recalibrated (version 2.o). The version, 1.0 or 2.0, is also specified in the method name (see table above).

The time in the returned table corresponds to the time for which the monthly sunspot number is forecasted (year/month), while the issue time is the time when the forecast is issued. The lead time is then the difference between these (it varies between -5 and 12 months as the official monthly sunspot number is only confirmed 6 months after the event, the forecast also covers the past 5 months).

The table displayed on the latest forecast page is always for version 2.0 and the corresponding six methods (SC2, SC2KF, CM2, CM2KF, ML2, ML2KF) and for all lead times (-5 to 12). Additionally, the latest forecast page provides a figure for each of the methods as shown for example below.

CM ML SC
KFCM KFML SCKF

The top menu navigation bar allows to browse through the products (across the different "Solar Weather" providers) as follows: