. 2019 Jul 18;10:809. doi: 10.3389/fpls.2019.00809

TABLE 1.

A detailed summary of the input variables for early-/mid-season SGB models.

Data source	Raw data	Model input variables	Variable name	Pearson correlation
Grower data
Yield records from 8 major almond growers in the Central Valley of California.	Planting year Historical yield Cultivar variety and composition	Age Previous two years’ yield Cultivar percentage	Age Pre 2Y Yld, Pre 1Y Yld CulP 1, …., CulP 20	−0.16^* 0.53^, 0.53^ −0.27∼0.25
Weather data
CA-BCM data with 270 m spatial resolution for years 1990–2016¹;	Monthly mean daily maximum temperature (Tmax); Monthly mean daily minimum temperature (Tmin); Monthly accumulative precipitation (PPT)	Current year monthly Tmax and Tmin from January to June², and PPT from January to March	Tmin, Tmax, PPT January Tmin, Tmax, PPT February Tmin, Tmax, PPT March Tmin, Tmax April Tmin, Tmax May Tmin, Tmax June	−0.06^, −0.06, −0.22^ −0.17^, 0.17^, −0.27^* 0.05^, 0.21^, −0.30^* 0.25^, 0.30^ 0.17^, 0.12^ 0.17^, 0.21^
		Previous year summer mean temperature averaged over July and August	Pre Tmean July–August	0.36^*
		Long-term mean seasonal Tmax, Tmin, PPT (averaged over 1990–2009 for each season³).	LT Tmin, Tmax, PPT January–March LT Tmin, Tmax, PPT April–June LT Tmin, Tmax, PPT July–September LT Tmin, Tmax, PPT October–December	0.26^, 0.50^, −0.58^* 0.43^, 0.60^, −0.61^* 0.38^, 0.49^, −0.57^* −0.31^, 0.51^, −0.58^*
CIMIS station data for years 2009–2017	Hourly temperature	Winter chilling portions calculated by the Dynamic Model	ChillP	−0.14^*
Remote sensing imagery
NAIP aerial imagery from 2016 with 0.6m resolution	NAIP RGB imagery acquired in 2016	2016 canopy cover percentage	CCP	0.20^*
Landsat satellite imagery with 30 m resolution years 2009–2017	Landsat multispectral imagery every 16 days	Previous year annual maximum NDVI and EVI; Current year June average EVI	Pre Max NDVI Pre Max EVI June Mean EVI	0.13^* 0.31^* 0.35^*

¹2017 Tmax, Tmin and PPT were collected from the original PRISM data archive. ²Variables used only for mid-season prediction models were shown in Italic. ³The four seasons are separated as (1) January–March, (2) April–June, (3) July–September, (4) October–December. ^*Represents for p < 0.05.