[[TOC]]

Sowing dates are simulated deterministically based on a set of rules depending on crop- and climate-specific characteristics. We assume that farmers base their timing of sowing on experiences with past precipitation and temperature conditions, with the intra-annual variability being especially important. The start of the growing period is assumed to be dependent either on the onset of the wet season or on the exceeding of a crop-specific temperature threshold for emergence. To validate the methodology, a global data set of observed monthly growing periods (MIRCA2000) is used.

We assumed that the timing of sowing is dependent on precipitation and temperature conditions, with the intra-annual variability of precipitation and temperature being especially important. Precipitation and temperature seasonality of each location are characterized by the annual variation coefficients for precipitation (CV _prec) and temperature (CV _temp), calculated from past monthly climate data (Fig. 1, src/crop/calc_seasonality.c
In order to simulate a reasonable global distribution of temperate and tropical regions, we assumed temperature seasonality if CV _temp exceeds 0.01. We assume precipitation seasonality if CV _prec exceeds 0.4.

Accordingly, four seasonality types can be distinguished (Fig.2) (see [src/crop/calc_seasonality.c] (https://github.com/PIK-LPJmL/LPJmL/blob/master/src/crop/calc_seasonality.c)):

1. no temperature and no precipitation seasonality
2. precipitation seasonality
3. temperature seasonality
4. temperature and precipitation seasonality

In situations with a combined temperature and precipitation seasonality, we additionally considered the mean temperature of the coldest month. If the mean temperature of the coldest month exceeded 10°C, we assumed absence of a cold season, i.e. the risk of occurrence of frost is negligible. Consequently, temperatures are high enough to sow year-round, therefore, precipitation seasonality is determining the timing of sowing. If the mean temperature of the coldest month is equal to or below 10°C, we assumed temperature seasonality to be determining the timing of sowing.

Figure 1: Procedure to determine seasonality type and sowing date.

Figure 2: Global distribution of seasonality types. For each seasonality type one example region is marked.

We applied specific rules per seasonality type to simulate sowing dates (Fig. 1).

In regions with no seasonality in precipitation and temperature conditions, crops can be sown at any moment and we assigned a default date as sowing date (1 January, for technical reasons).

In regions with precipitation seasonality, we assumed that farmers sow at the onset of the main wet season. The precipitation-to-potential-evapotranspiration ratio is used to characterize the wetness of months, as suggested by Thornthwaite (1948). Potential evaporation is calculated using the Priestley-Taylor equations with a value of 1.391 for the Priestley-Taylor coefficient. As a region may experience two or more wet seasons, the main wet season is identified by the largest sum of monthly precipitation-to-potential-evapotranspiration ratios of 4 consecutive months; 4 months were selected because the length of that period captures the length of the growing period of the majority of the simulated crops. Crops are sown at the first wet day in the main wet season of the simulation year i.e. with a daily precipitation higher than 0.1mm.

In regions with temperature seasonality, the onset of the growing period depends on temperature. Crop emergence is related to temperature, accordingly, sowing starts when daily average temperatures exceed a certain threshold (Table 1).

Table 1: Crop-specific temperature thresholds for sowing (Waha et al. 2012)

Crop	Temperature for emergence
Cassava	22
Groundnut	15
Maize	14
Millet	12
Pulses	10
Rice	18
Soybean	13
Spring rapeseed	5
Spring wheat	5
Sugar beet	8
Sugar cane	14
Sunflower	13
Winter rapeseed*	<17
Winter wheat*	<12

*Winter wheat and winter rapeseed are sown in autumn, as both crops have to be exposed to vernalizing temperatures.
Their base temperatures for emergence have been selected around the optimum vernalization temperatures.

The maps show:

• MIRCA2000 sowing dates (Portmann et al. 2008)
• Old sowing dates (Dec 2009)
• New sowing dates (May 2010)
• Difference map between MIRCA2000 and old sowing dates
• Difference map between MIRCA2000 and new sowing dates
• Difference map between old and new sowing date

Maps Wheat as representative crop for Temperate Cereals
Maps Millet as representative crop for Tropical Cereals
Maps Sugar-beet as representative crop for Temperate Roots
Maps Cassava as representative crop for Tropical Roots
Maps Rice
Maps Maize
Maps Pulses
Maps Soybean
Maps Groundnut
Maps Rapeseed

Only applied to crops, not to bio-energy plants.

Use flags in lpjml.conf:
NEW_SDATES: flag for the above described sowing date rules, default setting
READ_SDATES: read sowing dates from file sdate_combined.clm
FIX_SDATES: fix sowing dates after year specified in param.par

src/crop/calc_seasonality.c src/crop/sowing_season.c with READ_SDATES: src/crop/sowing_prescribe.c

Instead of calculating sowing dates according to the rules described above, one can also prescribe sowing dates from the global crop calendar MIRCA2000.

in data/biosx/LPJ/input_new:
sdate_combined.clm is the default and is a combination from the cell-specific and country-specific crop calendar.
sdate_maxarea.clm contains the sowing date from the main growing season with largest crop area.
sdate_combined_filled.clm, same as sdate_combined.clm but missing sowing dates are filled with average/maximum sowing date of the surrounding grid cells.

None of these are used for the paper, but the rainfed.sdate.bestseason.bin which contains the sowing date from the MIRCA2000 growing season corresponding
to the simulated sowing date. This means, that only one possible sowing date per grid cell is simulated, but there might be more. Especially in regions with
multiple cropping.

Please see README_prescribed_sowing_dates.txt and the Wiki page Input.

idea, concept, writing basic code and paper:
Katharina Waha, Lenny van Bussel, Christoph Müller, Alberte Bondeau

further code writing:
Jens Heinke, Christoph Müller

Bondeau, A., Smith, P.C., Zaehle, S., Schaphoff, S., Lucht, W., Cramer, W.,Gerten, D., 2007. Modelling the role of agriculture for the 20th century global terrestrial carbon balance. Global Change Biology. 13, 679-706.

Portmann, F.T., Siebert, S., Bauer, C.,Döll, P., 2008. Global dataset of monthly growing areas of 26 irrigated crops. Frankfurt Hydrology Paper. 6.

Portmann, F.T., Siebert, S.,Döll, P., 2010. MIRCA2000—Global monthly irrigated and rainfed crop areas around the year 2000: A new high-resolution data set for agricultural and hydrological modeling. Global Biogeochemical Cycles. 24, 1-24.

Input, Crop functional types, Vernalization

Waha, K., van Bussel, L.G.J., Müller, C.,Bondeau, A., 2012. Climate-driven simulation of global crop sowing dates. Global Ecology and Biogeography. 21, 247–259.

Journal version

Final draft