Prior to fermentation: total alcohol = natural alcohol plus potential alcohol arising from enrichment by adding sugar (or rectified concentrated grape must).
In fermented or bottled wine: total alcohol = actual alcohol plus potential alcohol content of any unfermented sugars.

Natural alcohol is the %vol that would be produced by fermentation of the juice/must if no enrichment were to take place. In Germany this is called 'available alcohol' which I think is a better term.
Under European Comission regulations applying to the United Kingdom the minimum permitted natural alcoholic strength of a must is 5 %vol (44 °Oechsle) for Table Wines, or 6 %vol (50 °Oechsle) for Regional (Table) Wines and Quality Wines. Other countries have higher levels for Quality Wine. I believe that the level for United Kingdom Quality Wine should be raised to 60 °Oechsle (7.7 %vol).
°Oechsle is simply specific gravity minus 1, then multiplied by 1000.

Actual alcohol is the %vol of the wine (in tank or bottle), measured by analytical techniques. This will be lower than total alcohol according to the amount of residual sugar in the wine.

For example: if the natural alcohol is 9.1 %vol and this must/juice is enriched by 2.3 %vol the total alcohol is 11.4 %vol; if the wine does not fully ferment out the actual alcohol will be slightly less than 11.4 %vol but the total alcoholic strength will still be 11.4 %vol.
Subsequently, if you add Süss-reserve at 11.4 %vol total alcoholic strength (84 °Oe) this 'dilution' will reduce the actual alcohol %vol but not the total alcohol; if the Süss-reserve is less than 84 °Oe then the total alcoholic strength will be reduced very slightly.

Regulations state that the actual alcoholic strength by volume (%vol) of wine in bottle shall be at lowest 8.5 %vol and not more than 15.0 %vol.

1) Firstly it is necessary to measure the temperature-corrected natural alcoholic strength of the must in °Oechsle (or Brix) at 20 °C, with a refractometer or hydrometer.
You can then look up the alcohol %vol in the UK Wine Standards Branch table; however, WSB figures are inaccurate (especially above 94 °Oe) so I have added (in brackets) accurate figures from German tables:-

  °Oechsle = alcohol %vol = Sugar g/l = % mas Sacch = Brix
      44            5.1 (5.13)         87              10.94
      45            5.3 (5.29)         90              11.18
      46            5.5 (5.45)         92              11.42
      47            5.6 (5.61)         95              11.66
      48            5.8 (5.76)         98              11.90
      49            5.9 (5.92)        100             12.14
      50            6.0 (6.08)        103             12.37
      51            6.2 (6.24)        106             12.61
      52            6.4 (6.39)        108             12.85
      53            6.6 (6.55)        111             13.08
      54            6.7 (6.70)        114             13.32
      55            6.9 (6.86)        116             13.55
      56            7.0 (7.02)        119             13.79
      57            7.2 (7.17)        122             14.02
      58            7.3 (7.33)        124             14.26
      59            7.5 (7.49)        127             14.49
      60              7.7 (7.65)        130               14.72
      61            7.8 (7.81)        132             14.96
      62            8.0 (7.96)        135             15.19
      63            8.1 (8.12)        138             15.42
      64            8.3 (8.27)        140             15.65
      65            8.4 (8.43)        143             15.88
      66            8.6 (8.59)        146             16.11
      67            8.7 (8.74)        148             16.34
      68            8.9 (8.90)        151             16.57
      69            9.1 (9.06)        154             16.80
      70            9.2 (9.22)        156             17.03
      71            9.4 (9.37)        159             17.26
      72            9.5 (9.53)        162             17.49
      73            9.7 (9.69)        164             17.72
      74            9.8 (9.84)        167             17.95
      75          10.0 (10.00)       170             18.18
      76          10.2 (10.16)       172             18.40
      77          10.3 (10.31)       175             18.63
      78          10.5 (10.47)       178             18.86
      79          10.6 (10.63)       180             19.08
      80          10.8 (10.79)       183             19.31
      81          10.9 (10.94)       186             19.53
      82          11.1 (11.10)       188             19.76
      83          11.3 (11.26)       191             19.98
      84           11.4 (11.41)       193               20.21
      85          11.6 (11.57)       196             20.43
      86          11.7 (11.73)       199             20.65
      87          11.9 (11.88)       201             20.88
      88          12.0 (12.04)       204             21.10
      89          12.2 (12.20)       207             21.32
      90          12.4 (12.35)       209             21.54
      91          12.5 (12.51)       212             21.77
      92          12.7 (12.67)       215             21.99
      93          12.8 (12.83)       217             22.21
      94          13.0 (12.98)       220             22.43
      95          13.0 (13.14)       223             22.65
      96          13.1 (13.30)       225             22.87
      97          13.3 (13.45)       228             23.09
      98          13.4 (13.61)       231             23.31
      99          13.6 (13.76)       233             23.53
     100          13.8 (13.92)       236            23.75
     101          13.9 (14.08)       239            23.96
     102          14.1 (14.24)       241            24.18
     103          14.2 (14.40)       244            24.40
     104          14.4 (14.55)       247            24.62
     105          14.5 (14.71)       249            24.83
     106          14.7 (14.87)       252            25.05
     107          14.8 (15.02)       255            25.27
     108          15.0 (15.18)       257            25.48
     109          15.2 (15.33)       260            25.70
     110          15.3 (15.49)       263            25.91

2) Decide on the total alcoholic strength of the enriched must, according to the type of wine you are making. The maximum enrichment allowed (from 2009 harvest onwards) is 3.0 %vol; unless a formal Request to increase the limit to 3.5 %vol is approved.

     For Sparkling wine, I aim for 10.5 %vol, as the secondary fermentation will add a further 1.5 %vol.

     For dry white wine, I aim for 10.5 or 11.0 %vol, according to the vine variety.

     For medium-dry or medium white wine, I aim for 11.5 %vol (the maximum permitted for enriched white wine), since this will be reduced later by the addition of Süss-reserve before bottling.

     For rosé, I aim for 11.5 or 12.0 %vol (the maximum permitted for enriched rosé wine).

     For red wine, I aim for 12.0 %vol (the maximum permitted for enriched red wine).

3) Calculate the weight of dry sugar required to achieve the desired total alcoholic strength; water may not be added to dissolve the sugar. Concentrated grape must or rectified concentrated grape must can also be used, but normally these are not used in UK.

Beet sugar or cane sugar may be used, beet sugar contains less pectin so is more likely to give a protein haze. Beet sugar is cheaper and hence is normally used, although cane sugar has more flavour.

a) For white wine or rosé, Wine Standards Board's figure of 16.5 grams of sugar per litre of must, to produce 1 % alcohol by volume, is slightly low; they admit it is conservative. German data is 16.85 g/l, Australia (according to Bryce Rankine) is 16.95 g/l and France (according to Peynaud) is 17 g/l.
According to Peynaud, cane sugar was used exclusively for white wines, although I doubt it is used in France today. The land around Champagne is a major sugar beet growing area.

Hence, for white wine or rosé: total grams of sugar to be added =
     total litres of must x (desired total alc %vol - natural alc %vol) x 16.85 g/l

Enrichment by 3.0 %vol (the maximum normally allowed) is achieved with 50.55 grams of sugar per litre of must.

It is essential to rouse continuously while sugar is being added, else the sugar will form a solid lump on the tank bottom.

b) For red wine, which is generally fermented at a higher temperature, 19 (according to Bryce Rankine) or 20 (according to Peynaud) grams of sugar per litre is needed to produce 1% alcohol by volume, because of loss of alcohol by evaporation during pumping over of the warm wine. It is the free-run wine that is most enriched; the press wine more or less keeps its initial strength. According to Peynaud, there is no distinction made between using beet or cane sugar for red wines.
Discussion in the UK has concluded that about 18.3 grams of sugar per litre of must, should be added in order to produce 1% alcohol by volume.

Hence, for red wine: total grams of sugar to be added =
     total litres of mash x forecast %-extraction x (12.0 alc %vol - natural alc %vol) x 18.3

Forecasting the %-extraction (% of must/juice that will eventually be pressed out from the de-stalked mash) is not easy and relies on experience. Each variety and vintage year is slightly different, according to the juiciness of the particular grapes. In my vineyard with the varieties Pinot Noir, Rondo, Regent, Dornfelder and Dunkelfelder, the overall average has been 83.3%, but individual wines have varied from 77 % (small dry berries) to 89 % (large juicy berries):-
  Pinot Noir (thin skins): from 82 % to 89 %, average 84.5 %
  Rondo is the most consistent: from 81.5 % to 84.5 %, average 83 %
  Regent/Dunkelfelder: from 77 % to 86.5 %, average 82.5 %
  Dornfelder: from 80 % to 89 %, average 83.5 %
This constitutes a possible error in enrichment of up to +-0.23 %vol.

Because red wine is fermented 'on the skins' it is not possible to use a rouser in the tank to dissolve the sugar, so an alternative method has to be used:-
- if an open-top tank is being used, the sugar can be added gradually day-by-day, simply tipping some on to the top of the mash just before the pomace cap is pushed down.
- or the technique of 'pumping over' can be employed. The sugar needed is tipped little by little into a small container fed by must from the tank. The resulting syrup is stirred continuously and pumped back to the top of the tank over the pomace cap. This is usually done in one step at the beginning of fermentation when the fermenting must is starting to get warm and the pomace cap is just forming. Sugar dissolves more easily in warm must.

Other sources of error in the enrichment process:-

1) Measurement of the volume of must:-

The % error for every 1 mm error in measurement of height in a vertical cylindrical tank = 1 divided by the height in mm; eg. for 1000 mm high must the error is 0.1 %. The diameter of the tank is less significant.

For tanks that do not have a flat base there is the additional problem of allowing for this.

The temperature of the must can lead to volume calculation errors.

A must that is not completely settled will contain suspended solids which will contribute a small volume calculation error, as would sludge in the bottom of the tank.

2) Initial measurement of grape-sugar (natural alcoholic strength):-

There are several methods of measuring the sugar content of a liquid, all require a representative, homogenous sample.

A hydrometer reading is affected by the temperature of the must and by the presence of material other than glucose/fructose in the must (which would increase the apparent must weight). The temperature effect can be eliminated by the use of a water bath, at 20 °C, to maintain a constant temperature for the measurements.

A refractometer reading is affected by the temperature of the must and by the presence of material other than glucose/fructose in the must (which could either increase or decrease the refractive index of the must and therefore give an apparent must weight that was higher or lower than the actual). The temperature effect is compensated for by a adding or subtracting a correction factor according to the measured temperature. The latest refractometers have built-in automatic temperature compensation.

Chemical analysis for reducing sugars is affected by other substances that reduce Cu^II to Cu^I (this would have the effect of increasing the apparent must weight). The analysis technique is straightforward.

Enzymic methods are affected by other substances that react with the enzyme or interfere with its action (leading to an apparent must weight that was higher or lower than the actual).

The only really accurate method, of determining sugars in grape must, is liquid chromatography. Here glucose and fructose are separated from other sugars, and from other compounds, in a liquid chromatography apparatus. This apparatus is prohibitively expensive to buy or operate.

3) Sugar addition:-

The scales used to weigh out the sugar may be inaccurate, leading to too little or too much alcohol being produced.

The sugar may not completely dissolve in the must, settling out in the sludge at the bottom of the tank, leading to too little alcohol being produced.

The types of sugar used for enrichment (granulated beet sugar, cane sugar, rectified concentrated grape must) will produce slightly different amounts of alcohol.

4) Natural variations in the fermentation:-

C₆H₁₂O₆ =  2 C₂H₅OH + 2 CO₂
Glucose/fructose = ethanol (ethyl alcohol) + carbon dioxide

180 g  = 2 x 46 g + 2 x 44 g
So, theoretically, 180 grams fructose/glucose = 92 grams alcohol

In practice the conversion of sugar (either natural grape sugar or added sucrose) to alcohol is not an exact process. It varies with the glucose/fructose balance in the grape, the fermentation temperature, yeast strain etc:-

The reaction sometimes does not go to completion, giving lower than expected alcohol.
The reaction can proceed by numerous routes to a variety of products, giving rise to lower than expected alcohol.
Other reactions may produce alcohol, giving rise to higher than expected levels.
The alcohol produced can be physically entrained, from the fermenting medium, by the action of carbon dioxide evolution and out-gassing. The level of entrainment will depend of the speed of fermentation and the fermentation temperature. This entrainment may also remove some water vapour, similarly affecting the volume of the ferment.

Bryce Rankine writes "100 grams of invert sugar produces between 45 and 48 grams of alcohol during complete fermentation". This variation for a wine which is expected to be 11.0 %vol gives a possible 'error' of +-0.35 %vol. Hence, at the time of enrichment, it is impossible to predict actual alcohol with accuracy greater than +-0.35 %vol, even without allowing for the other inaccuracies.

Comparing this practical result with the theoretical complete stoichiometric reaction you find that it is 88 % to 94 % of theoretical.
WSB's recommended chaptalisation (of 16.5 g/l sugar = 10 ml/l alcohol (1.0 %vol), compared with theoretical of 15.4 g/l = 10 ml/l) assumes a yield of 93.3 % of theoretical, so is always likely to result in low enrichment. The German figure of 16.85 g/l predicts a yield of 91.4 % of theoretical, which is close to the mid-point.