The moisture index is a measure of the soil moisture, and ranges from 0 (dry) to 1.0 (saturated). The water balance model used to compute the moisture index mimics the effects of precipitation and evaporation: for each week in turn it adds water to the soil from precipitation (up to the limits imposed by the soil characteristics), and takes it away by means of evaporation. Starting from the dry state, the entire model is run once with 52 weeks of data, and the output discarded. This lets the soil moisture value stabilise so that when the same sequence of precipitation and evaporation is applied again, the model has a realistic soil moisture store to begin with. The soil moisture index values from this second run are the ones used by BIOCLIM and GROCLIM.
The moisture index values are always based on weekly estimates. These are interpolated from the monthly climate estimates. In BIOCLIM, the calculated values are aggregated back into monthly values if the time step is set to months.
The moisture index is given by
(1-esoilb*store/maxstore) / (1-esoilb)
where store is the current store of water in the soil,
maxstore is the maximum soil water availability in mm and
soilb is a parameter that depends on soil type. Both BIOCLIM
and GROCLIM provide default values for soilb for a number of
soil types. These soil types are
.bcp file and the species
profile are affected in the same way by the approximation of soil
characteristics, the effects are minimal. Providing you use the same
soil moisture settings for all the BIOCLIM runs, they should compare
sensibly with each other, even if the soil moisture parameters are not
as exact as they could be.
In addition, compared to a plant growth process model using actual weekly or daily rainfall data, BIOCLIM is relatively insensitive to the soil moisture characteristics. A model that is driven with actual rainfall data (which is quite variable) is much more sensitive to soil moisture characteristics than is BIOCLIM, which deals with long-term monthly or weekly averages.
If you're really worried about the effect of the soil moisture characteristics, you can always run BIOCLIM on the same data but with different soil moisture values and see how the results compare.
For sites with very shallow and or rocky soils, you may find the results of the moisture index calculations counter-intuitive. These sites frequently have low plant available water and are constrained in their growth potential by moisture availability, but the monthly and even annual GROCLIM moisture index values are frequently rated as 1. The problem is not with the algorithms but is due to the need to aggregate rainfall into weekly chunks for the purposes of constraining the amount of data and calculations required.
This means that every week has (unrealistically) some input to the water balance calculations, which frequently stocks up soils with a low storage capacity, without allowing them to dry to a more realistic state of low water volume and potential that reflects moisture stress on plants. The early uses of GROCLIM were in an agricultural context in which relatively deep soils with a high moisture holding capacity were emphasised and in this situation the mismatch between the real distribution of rainfall and a simplified weekly stock up is less critical. That is, larger falls at less frequent intervals than weekly do not fill up the store more frequently than weekly inputs. Similarly in low rainfall environments this over-frequent filling of the soil water store is less likely to be a problem simply because rain is an infrequent event or the actual rainfall intensity is frequently low.
The quarterly parameters are not aligned to any calendar quarters. BIOCLIM's definition of a quarter is any 13 consecutive weeks, (or any consecutive 3 months if running with a monthly time step). For example, the driest quarter will be the 13 consecutive weeks that are drier than any other set of 13 consecutive weeks.