TRENDY-v9 (2020)
CRUJRA v2.1 forcing

1. Source Dataset
2. ORCHIDEE forcing
3. Interpolation to different resolution
4. Testing CDO interpolation from 0.5° to 2°

---

1. Source Dataset

• Based on : JRA-55 and CRU TS v4.04
• Period : 1901-2019
• Resolution : 0.5° x 0.5° (land surface excluding Antarctic)
• Longitudes : -179.75, -179.25 .. 179.75
• Latitudes : -89.75, -89.25 .. 89.75
• Time frequency : 6-hourly
• First time-step : at 00:00
• Calendar : 365d (noleap)
• 10 atmospheric variables :

Variable	Units	Long name
tmp	Degrees Kelvin	Temperature at 2m
tmin	Degrees Kelvin	Minimum temperature at 2m
tmax	Degrees Kelvin	Maximum temperature at 2m
pres	Pa	Pressure
spfh	kg/kg	Specific humidity
vgrd	m/s	Meridional component of wind speed
ugrd	m/s	Zonal component of wind speed
dswrf	J/m2	Downward solar radiation flux
dlwrf	W/m2	Downward long wave radiation flux
pre	mm/6h	Total precipitation

---

2. ORCHIDEE forcing

ORCHIDEE Variable	Conversion from CRUJRA	Units	Time shift	Cell method
Tair	tmp	K	1 time-step backward, so that the first data entry is at 06:00 (for compatibility with old driver)	timeplussix: instantaneous
PSurf	pres	Pa		timeplussix: instantaneous
Qair	spfh	kg/kg		timeplussix: instantaneous
Wind_N	vgrd	m/s		timeplussix: instantaneous
Wind_E	ugrd	m/s		timeplussix: instantaneous
Tmin	tmin	K	none	time: min(start)
Tmax	tmax	K		time: max(start)
SWdown	dswrf / (6*60*60)	W/m^2		time: mean(start)
LWdown	dlwrf	W/m^2		time: mean(start)
Rainf	pre / (6*60*60), if tmp >= 273.15 0, if tmp < 0	kg/m^2/s		time: mean(start)
Snowf	0, if tmp >= 273.15 pre / (6*60*60), if tmp < 0	kg/m^2/s		time: mean(start)

---

3. Interpolation to different resolution

ORCHIDEE Variable	Interpolation	Special treatment
Tair	bilinear + nearest neighbor if nan
PSurf	bilinear + nearest neighbor if nan
Qair	bilinear + nearest neighbor if nan	spfh → rh → interpolation → spfh (see equations below)
Wind_N	bilinear + nearest neighbor if nan
Wind_E	bilinear + nearest neighbor if nan
Tmin	bilinear + nearest neighbor if nan
Tmax	bilinear + nearest neighbor if nan
SWdown	bilinear + nearest neighbor if nan
LWdown	bilinear + nearest neighbor if nan
Rainf	nearest neighbor	pre → Rainf (if tmp >= 0) → interpolation
Snowf	nearest neighbor	pre → Snowf (if tmp < 0) → interpolation

Saturated water vapor pressure (\(p_{ws}\), Pa) at given temperature (\(t\), °C):

\( p_{ws}(t) = 611.2 \cdot \exp\left(\frac{\large{17.67 \cdot t}}{\large{243.5 + t}}\right) \)

Specific humidity (\(SH\), g/g) to relative humidity (\(RH\), %) conversion at given temperature (\(t\), °C) and pressure (\(p\), Pa):

\(RH = \frac{\large{SH \cdot p \cdot 100}}{\large{p_{ws}(t) \cdot (SH \cdot 0.378 + 0.622)}} \)

Relative humidity (\(RH\), %) to specific humidity (\(SH\), g/g) conversion at given temperature (\(t\), °C) and pressure (\(p\), Pa):

\(p_w = RH/100 \cdot p_{ws}(t)\)

\(SH = \frac{\large{0.622 \cdot p_w}}{\large{p - 0.378 \cdot p_w}}\)

(humidity conversion adapted from https://onlineconversion.vbulletin.net/forum/main-forums/convert-and-calculate/322-relative-humidity-absolute-humidity)

---

4. Testing CDO interpolation from 0.5° to 2°

CDO interpolation methods tested:

• nn (nearest neighbor)
• bil (bilinear)
• bilnn (bilinear + nearest neighbor where bilinear gives nan)
• con (conservative)

tmptmintmaxpresrhvgrdugrddswrfdlwrfpre

orig nn bil bilnn con

anm jan ts

Show Data