Divergent climate feedbacks in the growing period and the dormancy period to 1 sowing date shift of winter wheat in the North China Plain 2

s: The land cover and management changes have strong feedbacks to climate 18 through surface biophysical and biochemical processes. Agricultural phenology dynamic 19 exerted measurable impacts on land surface properties, biophysical process and climate 20 feedback in particular times at local/regional scale. But the responses of climate feedback 21 through surface biophysical process to sowing date shift in the winter wheat ecosystem 22 have been overlooked, especially at winter dormancy period. Considering the large 23 https://doi.org/10.5194/bg-2020-388 Preprint. Discussion started: 20 November 2020 c © Author(s) 2020. CC BY 4.0 License.

The NCP is the largest winter wheat production region in China, including Hebei,

133
The quality-controlled meteorological data, including air temperature (Ta), 134 precipitation (P), atmosphere pressure, relative humidity, and wind speed was obtained  Ta means air temperature, and P means precipitation. The meteorology conditions were also synchronously measured during flux 159 observation ( Table 2). The measurement included Ta, P, atmosphere pressure, relative 160 humidity, wind speed, and sunshine. These data was the inputs of the model. According

172
The phenology information was manually recorded and available in the period of  (Table 3). re-greening period is larger, and harvest period is relatively stable.

180
For the past 30 years, winter wheat phenology at some stations showed a significant 181 linear trend (Table 4). The sowing and germination periods were significantly delayed in 182 4 out of 10 stations, and the trend in the dormant and re-greening period was not obvious.

183
Winter wheat matured significantly earlier at five stations. Generally, the autumn and    as Rn, LH, and SH, which was used to explain the climate feedback mechanism.   Wheat LAI curves for the two sowing dates were obviously not overlapped (Fig.2a).

302
The LAI in the EP scenario was larger with earlier development. With the sowing in the 303 LP scenario, LAI difference between the two scenarios gradually narrowed until the 304 spring of the next year when the disparity increased again (Fig.3a). The LAI difference 305 between two scenarios had a valley after the reproductive period. With the approaching 306 of harvest, the difference gradually decreased to 0.

307
The LAI difference of winter wheat in two scenarios is mainly attributed to the 308 difference in the accumulation of organic matter. In the EP scenario, earlier sowing 309 means advanced assimilation process and better temperature conditions, more little to the LAI difference between the two scenarios.

316
The most obvious disparity in Tc between two scenarios occurred in the period when 317 wheat had been sown in the EP but hadn't in the LP (Fig.2b). The development of early 318 sown winter wheat resulted in higher Tc, with a peak of up to 0.6 K. The growth of wheat 319 in the LP sharply reduced the warming effect in EP, and eventually the EP scenario had 320 lower temperatures (-0.2K) before entering the dormancy period. The temperature change 321 process during this period was relatively consistent across the selected stations (Fig.3b).

322
Another special period is the dormancy period, when EP had higher Tc than LP with 323 average of 0.05 K (Fig.3b). With the start of the re-greening period, the EP Tc was February) and active growth periods (other wheat development period with active 352 physiological activity), but with both positive LAI difference (Fig.3). In this section, 353 surface energy balance was used to explain the response of Tc to sowing date.

354
The flux anomalies of Rn, LH, SH and R were shown in Fig.4a less SH was partitioned. Bigger anomaly of SH was happened in the initial and dormant 361 stages. R anomaly fluctuated obviously only in the initial phase.

362
The contributions of surface energy balance components to Tc were shown in Fig.4b.

363
Stronger radiation absorption provided more energy for the thermal motion of air and 364 causing positive Tc differences of EP-LP. Correspondingly, higher distribution into LH, 365 SH, and R was conducive to cooling Tc. Therefore, positive LH and SH differences of  Low soil water content also contributed to the high surface albedo (Seneviratne et al.

456
2010) (Fig.6b). With the decrease of surface soil moisture, surface albedo increased in 457 winter, which explained why albedo in the winter was higher than that in the growth 458 period. The increase in soil reflectivity caused by soil drying enhanced the role of low 459 winter wheat reflectivity in surface albedo, the albedo disparity between the two 460 scenarios increased in winter, so the albedo-radiative mechanism strengthened. Low soil 461 moisture also contributed to the disparity in warming effect between EP and LP during 462 dormancy period (Fig.6b)   (1) Earlier sowing date of winter wheat had higher LAI than later sowing date.

498
(2) The Tc disparity between EP and LP is divided into two periods: warming effect 499 in the dormancy period, and cooling effect in the active growth period.

500
(3) Surface energy balance can interpret the climate feedback mechanism of sowing 501 date shift, that is, the dominated role of albedo-radiative process in the dormancy period 502 is surpassed by LH partitioning-non-radiative process in the growth period.

503
(4) The responses of LAI and Tc to sowing date at station scale were divergent: 504 controlled by Ta in the dormancy period, and influenced by P and Ta in the growth period.