Geophysical Research Letters
Supporting Information for
Potential impacts of wintertime soil-moisture anomalies from agricultural irrigation at low
latitudes on regional and global climates
Hao-Wei Wey1, Min-Hui Lo1, Shih-Yu Lee2, Jin-Yi Yu3, Huang-Hsiung Hsu2
1Department of Atmospheric Sciences, National Taiwan University, Taipei, Taiwan, 2Research Center for
Environmental Changes, Academia Sinica, Taipei, Taiwan, 3Department of Earth System Science, University of
California Irvine, Irvine, CA, United States
Contents of this file
Text S1 to S4
Figures S1 to S8
Introduction
The supporting information provides detailed descriptions for the following issues.
Some details about the methodology of irrigation in the simulation are in Text S1. Comparisons
between CESM’s ET with AVHRR-based and FLUXNET-based ET are in Text S2 and Figure S1.
The simulated local impacts of irrigation are in Text S3 and Figures S2 and S3. Additional data
from ERA-Interim and CCSM4 last millennium simulations are used to support our findings on
the teleconnection responses to the interhemispheric temperature gradient in India Ocean
sector, and the method and results are in Text S4 and Figures S4 to S8.
Text S1
Previous studies have investigated the source of irrigation water [e.g., Siebert et al., 2010].
Siebert et al. [2010] implied that in South Asia, the portion of irrigation coming from
groundwater is about 50%, so we chose 50% of irrigation are from groundwater. However, we
do not have available data about how much of them are from confined and how much of them
are from unconfined aquifer, so we arbitrarily take half of them to come from unconfined
aquifer. As only unconfined aquifer is simulated in CESM, this is how we came up with the 1/4
estimate.
In the simulation, no water is removed from rivers, as originally the model does not evaporate
water from them, either. The model does evaporate water from lakes, and we ignore this part
as they are not significant in this region. As for the irrigation rates, they do vary from month to
1
month. The irrigation amount data we used are monthly, and we distribute them evenly to get
daily rates. Actually, the irrigation is applied at every time step to mimic the flooding behavior
of irrigation as previous studies have done (Lo and Famiglietti 2013; Lo et al. 2013) Previous
study also tested the sensitivity to the timing of irrigation, and they found no large differences
among simulations using different timing of irrigation (Sacks et al. 2009).
Text S2
The performance of land surface model (LSM) simulations in the CESM has been evaluated in
previous studies (Lawrence et al. 2011, 2012). Lawrence et al. [2011] noticed that the
parameterization scheme for evapotranspiration has been modified; partitioning of ET into
transpiration, ground evaporation, and canopy evaporation in version 4 of the Community Land
Model (CLM4) has been improved. And in Lawrence et al. [2012], they further assessed the LSM
simulations in the Community Climate System Model, version 4 (CCSM4). Using also the
FLUXNET-based latent heat flux data for comparison, they showed that the global centered
RMSE is lower in CCSM4, indicating the better simulation of the phase and amplitude of the
latent heat flux annual cycle (see their Table 1). Although there was excessive latent heat over
tropical and midlatitude regions, the errors were much lower in offline LSM simulations,
indicating the contribution of the errors from the coupled system. For selected regions, the
model also produced good results in terms of both phase and amplitude (see their Figure 5). We
also conduct the same comparisons for CESM1 and show in Figure S1 the results from top 10
river basins with largest areas. The results show that the model simulated evapotranspiration is
quite comparable to the observations and careful inspections on the comparisons are
worthwhile. The model produces rather good results in winter and the overestimate generally
happens in summer, making the underestimate in IGP in winter more unique. Besides, there is
also no river basin having the bimodal characteristics as we see in IGP. What we did in the
current study is to look into the possibly missed or improperly represented physical processes,
and find that the absence of irrigation in the simulation might be a candidate for the reason of
underestimation of evapotranspiration over IGP during the wintertime.
Text S3
Figure S2 provides the JFM local simulated responses over South Asia to the irrigation water
applied. To identify the local climate responses to South Asia irrigation, we compared and show
in Figure S2 the simulation differences between the CTR and IRR simulations in JFM. Over the
South Asia, the irrigation increases the soil water in the top 10 cm of the soil layer (Figure S2a)
and the latent heat fluxes into the atmosphere (Figure S2b). The mean latent heat release in
the IGP changed from 25.92 W/m2 in the CTR simulation to 37.74 W/m2 in the IRR simulation.
The latter value is reasonably close to the value of the observed latent heat flux shown in Figure
1c. We note that the largest increases of soil water and latent heat fluxes occur in the heavily
irrigated regions (cf. Figures S2a and S2b to Figure 1a). By contrast, low-level cloud cover
increases downwind of the irrigated regions (Figure S2c). The increased cloudiness decreases
the downward solar radiation at the surface (Figure S2d) to cause a surface cooling that spreads
throughout the entire Indian subcontinent (Figure S2e).
In addition, in this study, we show the differences of soil water in the top 10 cm, but the positive
soil moisture anomalies are observed at the deeper soil layer as well (Figure S3).
2
Text S4
We provide some examinations on the results to strengthen the findings in this study by using
the ERA-Interim reanalysis data and CMIP5 Community Climate System Model version 4
(CCSM4) last millennium simulations to analyze the impacts of South Asian winter monsoon
variability from Indian subcontinent surface temperature on northeastern Pacific.
There are approximately 30 and 1000 years of data in ERA-Interim reanalysis and CCSM4 last
millennium simulations, respectively. We define an Indian interhemispheric temperature
gradient, ΔT=TSH-TNH, which is the temperature difference between SH and NH part of Indian
basin (TSH is 0-30°S, 40-100°E and TNH is 0-30°N, 40-100°E; see Figure S4), to represent the
irrigation induced surface temperature gradient between the land and the ocean. Then we
choose the winters with largest and smallest ΔT (ΔT larger and smaller than 1 (2) standard
deviations for ERA-Interim (last millennium simulations)) and take their differences of sea level
pressure and upper level zonal wind velocity (see below Figures S5-S8). It is known that ENSO
has prominent influences on extratropical circulation [e.g., Horel and Wallace, 1981; Wallace
and Gutzler, 1981]. But meanwhile, it also influences the Indian Ocean simultaneously. To
remove the influences of ENSO, we should only composite years with neutral phase of ENSO.
Therefore, we only choose years with absolute values of normalized Niño3.4 index smaller than
1(0.1) for ERA-Interim (last millennium simulations) in our composites. In Figures S5-S8, it can
be seen that in both the reanalysis and last millennium simulations, Aleutian Low is deepened
and the subtropical jet over northeastern Pacific is intensified when ΔT is larger.
The analysis from ERA reanalysis and CCSM4 last millennium simulations clearly illustrate the
potential linkage of Indian subcontinent surface temperature and thus South Asian winter
monsoon and northeastern Pacific climate.
3
References
Horel, J. D., and J. M. Wallace, 1981: Planetary-scale phenomena associated with the Southern
Oscillation. Mon. Wea. Rev., 109, 813–829, doi:10.1175/15200493(1981)109<0813:PSAPAW>2.0.CO;2.
Lawrence, D. M., and Coauthors, 2011: Parameterization improvements and functional and
structural advances in Version 4 of the Community Land Model. J. Adv. Model. Earth Syst.,
3, 1–28, doi:10.1029/2011MS000045.
Lawrence, D. M., K. W. Oleson, M. G. Flanner, C. G. Fletcher, P. J. Lawrence, S. Levis, S. C.
Swenson, and G. B. Bonan, 2012: The CCSM4 land simulation, 1850-2005: Assessment of
surface climate and new capabilities. J. Clim., 25, 2240–2260, doi:10.1175/JCLI-D-1100103.1.
Lo, M. H., and J. S. Famiglietti, 2013: Irrigation in California’s Central Valley strengthens the
southwestern U.S. water cycle. Geophys. Res. Lett., 40, 301–306, doi:10.1002/grl.50108.
——, C. M. Wu, H. Y. Ma, and J. S. Famiglietti, 2013: The response of coastal stratocumulus
clouds to agricultural irrigation in California. J. Geophys. Res. Atmos., 118, 6044–6051,
doi:10.1002/jgrd.50516.
Sacks, W. J., B. I. Cook, N. Buenning, S. Levis, and J. H. Helkowski, 2009: Effects of global
irrigation on the near-surface climate. Clim. Dyn., 33, 159–175, doi:10.1007/s00382-0080445-z.
Siebert, S., J. Burke, J. M. Faures, K. Frenken, J. Hoogeveen, P. Döll, and F. T. Portmann, 2010:
Groundwater use for irrigation - A global inventory. Hydrol. Earth Syst. Sci., 14, 1863–1880,
doi:10.5194/hess-14-1863-2010.
Wallace, J. M., and D. S. Gutzler, 1981: Teleconnections in the Geopotential Height Field during
the Northern Hemisphere Winter. Mon. Weather Rev., 109, 784–812, doi:10.1175/15200493(1981)109<0784:TITGHF>2.0.CO;2.
4
Fig. S1. Comparisons of CESM-simulated latent heat flux compared with observations (W/m2)
5
Fig. S2. Simulated differences (IRR-CTR) in (a) 10-cm soil water (kg/m2); (b) latent heat flux
(W/m2); (c) low cloud cover; (d) surface-downwelling shortwave flux (W/m2); and (e) reference
height temperature (K) (dotted: p < 0.1).
6
Fig. S3. Simulated differences (IRR-CTR) of soil water content (kg/m2) in the eighth layer of soil
at around 1 meter depth
7
Fig. S4. Map showing the region where we take the mean surface temperatures TSH (0-30°S,
40-100°E) and TNH (0-30°N, 40-100°E)
8
Fig. S5. Composite 200-hPa zonal wind difference (ERA-Interim reanalysis) (m/s)
9
Fig. S6. Composite 200-hPa zonal wind difference (Last Millennium simulation) (m/s)
10
Fig. S7. Composite sea level pressure difference (ERA-Interim reanalysis) (Pa)
11
Fig. S8. Composite sea level pressure difference (Last Millennium simulation) (Pa)
12