- Email: [email protected]
Cotton crop phenology in a new temperature regime
Ecological Modelling 285 (2014) 22–29
Contents lists available at ScienceDirect
Ecological Modelling journal homepage: www.elsevier.com/locate/ecolmodel
Cotton crop phenology in a new temperature regime Qunying Luo a,∗ , Michael Bange b , Loretta Clancy b a b
University of Technology, Sydney, PO Box 123, Broadway, NSW 2007, Australia CSIRO Division of Plant Industry, Locked Bag 59, Narrabri, NSW 2390, Australia
a r t i c l e
i n f o
Article history: Received 1 February 2014 Received in revised form 15 April 2014 Accepted 16 April 2014 Available online 13 May 2014 Keywords: Temperature increase Cotton phenology Heat stress Cold shock
a b s t r a c t The daily outputs of the CSIRO Conformal Cubic Atmospheric Model driven by four general circulation models (GCMs) were used by a stochastic weather generator, LARS-WG, to construct four local climate change scenarios (CCSs) at nine key cotton production areas in eastern Australia. These CCSs were then linked to daily temperature-driven models of cotton phenology to examine the magnitude of the effects of increased temperature on the initiation and duration of key crop phenophases and on the occurrence of heat stress and cold shocks during the growing season. The results show that when using 1st Oct. as sowing time (1) the timing of emergence, 1st square, 1st flower and 1st open boll advanced 1–9, 4–13, 5–14 and 8–16 days, respectively, for the period centred on 2030 compared to baseline; (2) when crops were planted 10 days earlier, the emergence stage occurred earlier in most of the locations while other phenological events changed slightly (∼1 day) in comparison with 1st Oct. sowing; when crops were planted 10 days later, all these events were generally delayed (∼1.5 days) in comparison with 1st Oct. sowing; (3) the timing of the last effective square, last effective flower and last harvestable boll were delayed 7–12, 6–9 and 3–9 days, respectively, across locations (except St George) and GCMs; (4) the fruit development period increased up to 2–3 weeks; (5) the number of hot days increased across all locations and growing season (GS) months except May with the warmer months (Dec., Jan. and Feb.) and locations increased more; and (6) the number of cold shocks decreased or remained the same across locations and GS months except Jan. and Feb. with cold months and places decreased more. The results show that there will be less impact of cold temperatures on earlier growth and potentially a longer growing season that can improve crop yield. However, there will be more incidences of hot days impacting growth, and more rapid crop development in late phenological stages (especially during boll filling) that may limit the opportunities associated with increases in growing season length without adjustments in management. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Ambient temperature (T) is one of the major environment variables affecting the growth, development and yields of crops, especially the rate of development (Luo, 2011). Crops have basic requirements for T to complete a specific phenophase or the whole life cycle. On the other hand, extremely high and low temperatures (Ts) can have detrimental effects on crop growth, development and yield, particularly at critical phenophases. While cotton is morphologically indeterminate, the rate of many phenological processes such as germination, floral initiation and development of fruiting bodies is controlled by T (Hearn and Constable, 1984). Average daily T also plays an important role in determining the earliest date of sowing, defining season length which can both influence yield potential and cotton fibre quality (Bange et al., 2008a; Bauer et al.,
∗ Corresponding author. Tel.: +61 0410 186 352; fax: +61 2 9514 4079. E-mail address: [email protected] (Q. Luo). http://dx.doi.org/10.1016/j.ecolmodel.2014.04.018 0304-3800/© 2014 Elsevier B.V. All rights reserved.
2000; Dong et al., 2006), and determining where cotton can be produced sustainably. Generally the longer the growing season (GS) the greater the potential for higher lint yields (Bange and Milroy, 2004a; Stiller et al., 2004). In Australian cotton production systems, T requirements for the development of cotton are described by the accumulation of degree days calibrated with a base T of 12 ◦ C (Constable and Shaw, 1988). Other cotton systems elsewhere (i.e. America) have used a base T closer to 15 ◦ C (Robertson et al., 2007). Throughout most Australian production systems, minimum temperature (Tmin ) (≤11 ◦ C) experienced early in the cotton season can cause delays in development, reduction in growth, and sometimes chilling injury (Constable et al., 1976; Bange and Milroy, 2004b). As the season progresses maximum Ts (>35 ◦ C, heat stress) are commonplace throughout cotton production regions and may adversely affect growth and development thus affecting water use, yield and fibre quality (Hodges et al., 1993). At the end of a cotton season cooler or cold T will influence the timing of crop maturity and impact on effectiveness of chemical harvest aids, both again directly affecting yield and quality. Cotton has an optimal thermal
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
23
Fig. 1. Cotton production locations in Eastern Australia used in this study.
kinetic window of 23–32 ◦ C in which metabolic activity is most efficient (Burke et al., 1988; Conaty et al., 2012). High T affects reproductive processes of cotton thus affecting yield (Reddy et al., 2000, 2005). Effects may be attributed to impacts on pollen viability (Burke et al., 2004), fertilisation efficiency (Snider et al., 2009), boll size and seed number (Pettigrew, 2008), and fruit shedding (Hodges et al., 1993; Reddy et al., 1992b). Kakani et al. (2005) reported that instantaneous air T above 32 ◦ C reduces cotton pollen viability, and T above 29 ◦ C reduces pollen tube elongation. Reddy et al. (2005) demonstrated that the maximum growth rate per boll occurred at 25–26 ◦ C and decreased at higher Ts while boll harvest index was highest at 28 ◦ C, declining thereafter and reaching zero boll harvest index at 33–34 ◦ C. Boll size was largest when Ts less than 20 ◦ C. Percent boll set, boll number, boll filling period, rate of boll growth, boll size and yield all decreased when mean T reached 30 ◦ C (Reddy et al., 2005). T of 40/32 ◦ C led to zero boll yield (Reddy et al., 1992a,b). Heat stress can result in boll cavitation and reduces cotton lint yield (Bange et al., 2008b). From a fibre quality perspective sustained increases in T will reduce fibre length (Gipson and Joham, 1969) and increase micronaire (Gipson and Joham, 1968; Bange et al., 2010), and sometimes negatively impacting final fibre quality. Cotton exposed to low Ts takes longer to develop, and accumulates biomass at a slower rate (Gipson, 1974; Mauney, 1986). Cold T responses (below 10 ◦ C) in early post-emergent seedlings can permanently arrest growth and development (Christiansen, 1967; Christiansen and Thomas, 1969). Bange and Milroy (2004b) also found that negative effects on development existed when plants were exposed to at least 10 nights at 10 ◦ C, or for 5 nights at 5 ◦ C on post-emergent cotton. In Australia, to account for early season effects of cold T on cotton development, a ‘cold shock’ effect is applied to degree days (DDs) accumulation. A ‘cold shock’ is defined
as an event when the daily minimum T falls to 11 ◦ C or less, which assumed to cause chilling injury and thus delays development. Late in the season cold Ts (≤2 ◦ C) can signal the end of the GS. These low T may force bolls to open affecting lint fibre quality (colour and maturity), and will severely impede the efficacy of chemical harvest aids (optimal T ∼18 ◦ C) to remove leaf from the plant prior to harvest (Bange et al., 2009). The optimum sowing time aims to reduce the incidences of ‘cold shocks’, while in-season management aims to ensure that fruit has adequate time to mature and that harvest aids are applied prior to the onset of cold T. Increase in T will change crop phenology, including the start and duration of phenophases and the probability and degree of extreme climate events such as heat stress and cold shocks. These changes may have significant implications for cotton lint yield and fibre quality. Hence this work aims to understand local climate change in Australian cotton production regions, quantify its impacts on cotton crop phenology and on the occurrence of heat stress and cold shocks in cotton GS. This work will support more complex studies and analyses into climate change on cotton production and assist in identifying effective adaptation options to deal with the risk and to take the advantage of climate change. 2. Materials and methods 2.1. Study locations This study focused on nine major cotton production areas in Queensland (Emerald, Dalby, St George and Goondiwindi) and New South Wales (Moree, Bourke, Narrabri, Warren and Hillston) in Australia (Fig. 1). These cotton production areas also represent different growing environments: with Emerald, Bourke and St George being classified as being hot (annual mean T: 20.21–22.61 ◦ C);
24
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
Dalby, Goondiwindi, Moree, Narrabri and Warren as being mild (annual mean T: 18.21–19.91 ◦ C); and Hillston as cool (annual mean T of 17.58 ◦ C). Classification of these climates for cotton production was based on the analysis of McMahon and Low (1972) using growing DD. These different environments have resulted in adoption of different crop management practices and cultivars. 2.2. Local climate change and climate change scenarios Dynamic downscaling is one of the major downscaling approaches for translating coarser spatial resolution climate change information to finer scales (Luo & Yu, 2012). In this study, the outputs of the CSIRO Conformal Cubic Atmospheric Model (CCAM), a dynamic downscaling approach, for baseline (1980–1999) and future period (2020–2039) were used by a stochastic weather generator, LARS-WG, to derive local T change including changes in the mean (i.e., Tmax , Tmin ) and in variability (i.e., mean T variability). For the performance of this weather generator, readers are directed to Online Resource 1. The CCAM model was driven by four GCMs specifically GFDL, CSIRO Mark 3.5, MPI, and MIROC, under the Special Report on Emission Scenarios (SRES, Nakicenovic and Swart, 2000) A2 scenario. The rationale of using the outputs of CCAM driven by four GCMs is that multi-model ensemble is intrinsically more useful and accurate than sets of projections produced with any single model (Barnston et al., 2003; Hagedorn et al., 2005; Weigel et al., 2008). More information on CCAM is provided by McGregor and Dix (2008) and Luo et al. (2010). The local T change information derived earlier for each location were then reapplied to the LARSWG along with the historical daily T data to construct long time series (100y) temperature scenarios including baseline and future scenarios for impact assessment. The period 1980–1999 is standard baseline period. We chose 2020–2039 as the future period as farmers are more interested in near timeframe climate change rather than many (i.e. 40) years later. 2.3. Cotton crop phenology and occurrence of extreme temperatures Cotton sowing in the Australian industry occurs around 1st Oct. (Bange et al., 2008a). In this analysis, we chose three sowing times: 21st Sep., 1st Oct. and 11th Oct. We then defined the GS as the period from these sowing dates until 21st May, 31st May and 10th Jun., respectively. For the mild and cool locations these sowing dates generally represent a range of sowing dates currently used. Cotton crop phenology stages considered in this study include the date of emergence, first square (flower bud), first flower, first open boll, last effective square (LES), last effective flower (LEF), and last harvestable boll (LHB). The DD targets corresponding to the first four stages are set to 80, 505, 777 and 1527 DD, respectively (Constable and Shaw, 1988). The growth targets used have been calibrated with hot days occurring during development. The last harvestable boll is associated with the occurrence of the first frost, which is defined as the Tmin of ≤2 ◦ C. The date of the last effective square, flower, and boll can be used to determine target ‘cutout’ dates that assist in ceasing production of new fruiting sites to allow crops (and all bolls) to be mature in time for harvest. The DD differences between the date of LHB and LEF (LHB period) and between the LEF and LES (square period) are set to 750 and 430 DD, respectively (Constable, 1991). The DD were derived using a base T of 12 ◦ C (Constable and Shaw, 1988): DD =
[(Tmax − 12) + (Tmin − 12)] 2
(1)
where Tmax and Tmin are daily maximum and minimum Ts, respectively. When Tmin < 12 ◦ C, Tmin = 12 (or (Tmin − 12) ≥ 0.0).
As mentioned in Section 1, to account for cold T on early season cotton development a ‘cold shock’ effect is applied to DD accumulation for predicting the first square and first flower. A ‘cold shock’ is defined as an event when daily Tmin falls to 11 ◦ C or less, and extends the DD accumulation by 5.2 DD for each cold shock event (Hearn and Constable, 1984). Extreme Ts considered in this study include the occurrence of heat stress and cold shock on a monthly basis within the GS (1 Oct.–31 May). A heat stress event occurs when daily Tmax is ≥35 ◦ C. To quantify the effects of climate change on crop phenology, the on-line DD Calculator and Last Effective Flower tool (http://cottassist.com.au), developed by CSIRO, were modified to be used with baseline (1980–1999) and future period (2020–2039) T scenarios for the study locations. A programme was edited to quantify the occurrence of cold shocks and heat stress on a monthly basis. Changes in cotton phenology and in the occurrence of heat stress and cold shocks were presented in multi-model ensemble means, which is the average derived from the four driving GCMs. The analysis also assumes that there are no effects of elevated CO2 on the rates of cotton development. Reddy et al. (1999) found that there is no direct effects of enhanced CO2 on phenology, however, a more recent studies by Yoon et al. (2009) at very high levels of CO2 (800 mol l−1 ) found that there was more rapid development to flowering at moderate daytime Ts (25 ◦ C day/15 ◦ C night). It should be noted that the occurrence of late phenological events and of cold shocks and heat stress is independent of sowing times. 3. Results 3.1. Temperature change Fig. 2 shows the multi-model ensemble mean changes between the periods centred on 2030 and 1989 in Tmax , Tmin , and Tmean for GS and for each location. Tmax (0.9–1.1 ◦ C), Tmin (1.1–1.3 ◦ C) and Tmean (1–1.2 ◦ C) all increased across locations with Tmin increasing the most (Fig. 2). The largest change in average T occurred in Dalby, associated with a considerable shift in both Tmax and Tmin . The smallest change in Tmean was at Hillston associated with the lowest changes in both Tmax and Tmin across locations. The largest (climate) model-to-model variation (error bar) associated with Tmax was found at St George while the least was found at Emerald and Hillston. The increase in Tmax is greater than 0.6 ◦ C for all models and locations. For Tmin , the least model-to-model difference was found at Emerald and the others were fairly similar. The increase in Tmin was greater than 1 ◦ C for all models and locations. 3.2. Changes in crop phenology 3.2.1. Normal sowing Fig. 3 shows the multi-model ensemble mean changes in key phenological events for the period centred on 2030 in comparison with the baseline period 1980–1999. T increase accelerated the phenological development of all phases across all locations (Fig. 3). Emergence, first square, first flower and first open boll individually were advanced 1–9, 2–8, 1–2 and 1–5 days, respectively (Fig. 3a). The largest changes were projected in the period from sowing to emergence, especially for the locations further south. In general the smallest effect across all phases was associated with the most northern growing region of Emerald while the largest change for all phases was associated with the most southern growing areas of Warren and Hillston. This translated into the smallest overall cumulative change for Emerald (mean change of 8 days) and greatest change for Hillston and Warren (mean change 16 days). For the end of season, the timing of the phenological stages of the LES, LEF and LHB were delayed 0.3–3, 1.8–6.8 and 2.5–9 days,
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
25
a
Emerald Dalby St George Goondiwindi Moree Bourke Narrabri Warren Hillston 0.6
0.8
1.0
1.2
1.4
Changes in Growing Season Maximum Temperature (oC)
b
Emerald Dalby St George Goondiwindi Moree Bourke Narrabri Warren Hillston
0.6
0.8
1.0
1.2
1.4
Changes in Growing Season Minimum Temperature (oC)
Emerald
c
Dalby St George Goondiwindi Moree Fig. 3. Cumulative contribution of each phenophases to cotton crop phenology in 2030: (a) earlier phenological stages associated with normal sowing (sown on the 1st of Oct.) in comparison with baseline and (b) later phenological stages.
Bourke Narrabri Warren Hillston 0.6
0.8
1.0
1.2
1.4
Changes in Growing Season Mean Temperature (oC) Fig. 2. Changes in growing season (1st Oct.–31st May) average temperatures for 2030: (a) GS Tmax , (b) GS Tmin and (c) GS Tmean . Horizontal line is the error bar, which is calculated as the standard deviation divided by root mean square of the sample size. The empty circle is the multi-model ensemble mean.
respectively, across all locations except at Warren where the LEF was unaffected and at St George where the LHB advanced slightly (Fig. 3b). The greatest variability in delays was associated with LHB (at T ≤ 2 ◦ C) with the largest changes projected for Warren (mean change 9 d), Dalby (mean change 7 d) and Narrabri (mean change
6 d). The smallest changes for LHB were for St George (mean change −0.3 d) and Bourke (mean change 2.5 d). Across all locations, LES and LEF were also later as the periods between LHB to LEF, and LEF to LES were shorter. Overall, this meant that the smallest overall delays in season end were for Bourke (mean change 6.5 d), Hillston (mean change 7.3 d) and St George (mean change 7.5 d). The projected longest delays in season end were for Dalby (mean change 11.8 d) followed by Moree (mean change 10 d). 3.2.2. Earlier sowing With earlier sowing (21st Sept.), emergence advanced by 1–3 d or was unchanged across locations except for St George where a 3 d delay was found. First square was advanced by 1–4 d in northern cotton production areas (with warmer climate) but was delayed ∼1 d in southern areas (with colder climate). In general there was
26
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
Table 1 Changes (in days) in early phenological events of an earlier (21 Sep.) and later sowing (11 Oct.) in comparison with 1 Oct. sowing in a changing climate. Locations
Emergence
1st square
1st flower
1st open boll
Earlier sowing Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
−1.5 −1.0 −1.5 −3.0 −2.3 0.0 3.0 −1.5 0.0
0.8 0.0 0.8 0.5 0.8 −2.0 −3.8 −0.5 −0.8
0.5 0.3 0.5 −0.5 0.3 0.5 −0.8 0.5 0.8
−0.5 −0.3 0.0 0.5 0.0 0.5 1.0 0.5 −0.3
Later sowing Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
−0.3 0.0 2.8 2.3 4.0 0.5 2.3 3.3 0.5
0.3 2.8 0.3 −1.0 −1.5 0.5 0.0 −1.5 −0.5
1.3 0.8 −0.5 −0.5 0.3 −0.5 −0.5 −0.5 0.5
−1.5 −0.8 0.0 −0.3 −0.3 1.0 0.5 0.0 −0.3
decreased or unchanged across locations and GS months except Jan. and Feb. where minor increase occurred in some of the locations especially NSW cotton production areas. Generally speaking, relatively larger decrease in the no. of cold shocks was found at colder/southern locations (i.e. Hillston & Warren) and vice versa for the warmer places (i.e. Emerald) in most of the months. The minor increase in the number of cold shocks in Jan. and Feb. at some of the locations especially NSW cotton production areas reflected the effects of changed T variability on this parameter as increased T variability can lead to both increase of hot days and cold shocks. Both spatial and temporal variability exist associated with these two extreme Ts. 4. Discussion 4.1. Local temperature change
little effect on the 1st open boll as advancement or delay was no more than 1 d (Table 1) 3.2.3. Later sowing With later sowing (11th Oct.) in comparison with 1st Oct. sowing, emergence delayed 1–4 d except at Hillston where a 0.3 d advancement was found. First square was advanced by 1–2 d in warmer locations while delayed 0.3–3 d in colder locations. First flower and 1st open boll were advanced or delayed in the range of 0–1.5 d depending on locations (Table 1). 3.3. Occurrence of extreme temperatures Table 2 shows the frequency of heat stress days (>35 ◦ C) and cold shocks (≤11 ◦ C) within GS months for the nine cotton production areas under investigation. In 2030, the number of heat stress days increased across all locations and across GS months except May where no change was found. Larger increase occurred in warmer months (Nov.–Mar.) and warmer locations (i.e. Emerald, Goondiwindi and St George). On the contrary, the number of cold shocks
Tmax , Tmean and Tmin increased 0.9–1.1 ◦ C, 1–1.2 ◦ C and 1.1–1.3 ◦ C, respectively (Fig. 2) across locations and GCMs for the period centred on 2030 under A2 emission scenario, which represents a medium–high emission scenario. Overall, increase in Tmin is larger than that of Tmax . This finding is in line with earlier climate change studies conducted in other countries (i.e. Karl et al., 1997). Even though all four GCMs projected an increase in Ts, model-tomodel differences exist with larger difference found at St George and smaller at Hillston and Emerald associated with Tmax . CSIRO and BOM (2007) reported an increase of 1–1.5 ◦ C (median projection) for the inner part of Australia under medium–high emission scenario. Our results are very close to the results reported by CSIRO and BOM (2007). Spatial variation exists in the increase of T indicating different management options and/or efforts are needed to tackle the effects of climate change in different cotton production regions. This analysis considered uncertainties due to climate modelto-model difference. The local climate change was derived from a dynamical downscaling approach: one of the mainstream downscaling approaches. The limitation of this analysis is that only one emission scenario (A2) was considered due to the large computing resources needed to undertake this type of downscaling approach. Nonetheless it represents a relatively high emission scenario with corresponding higher increase in T compared with other low emission scenarios. Therefore impact analysis and adaptation evaluation based this emission scenario should capture the upper range of the
Table 2 Changes in the frequency of heat stress events (>35 ◦ C) and cold shocks (≤11 ◦ C) within GS months (Oct.–May inclusive). Numbers in brackets are the frequency of extreme Ts under baseline climate conditions. Oct. Heat stress Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
0.5 (0.5) 0.7 (0.5) 0.8 (0.4) 1.8 (1.3) 0.8 (0.5) 1.4 (0.7) 1.8 (1.1) 1.5 (0.4) 2.9 (2.9)
Cold shocks Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
−5.1 (19.4) −6.2 (18.2) −5.3 (13.9) −3.9 (7.9) −3.1 (11.0) −2.6 (6.8) −2.7 (5.9) −3.0 (8.8) −0.6 (1.0)
Nov. 0.9 (1.8) 1.8 (2.2) 1.9 (2.5) 3.1 (5.9) 2.3 (2.4) 2.2 (2.6) 4.8 (4.5) 1.9 (1.3) 6.2 (6.9) −4.3 (8.4) −2.5 (4.9) −1.8 (3.1) −0.5 (1.0) −1.3 (2.9) −0.5 (1.2) −0.4 (0.8) −0.7 (1.6) 0.0 (0.0)
Dec. 1.6 (6.8) 3.7 (7.2) 5.0 (7.7) 4.4 (14.9) 4.6 (6.0) 5.7 (6.6) 5.3 (11.8) 3.0 (3.1) 5.8 (13.9) −1.7 (2.1) −0.6 (0.9) −0.4 (0.6) −0.1 (0.1) −0.3 (0.5) −0.1 (0.1) 0.0 (0.0) −0.1 (0.2) 0.0 (0.0)
Jan.
Feb.
3.3 (10.1) 2.1 (12.4) 2.6 (11.6) 0.0 (19.3) 3.4 (9.3) 3.4 (10.0) 0.5 (15.6) 2.9 (3.4) 5.4 (12.6)
2.0 (7.1) 2.6 (6.7) 2.9 (6.9) 2.3 (12.6) 3.9 (5.6) 3.7 (5.7) 2.4 (9.4) 1.8 (2.0) 3.1 (6.8)
0.3 (0.5) 0.1 (0.1) 0.0 (0.1) 0.1 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.1) 0.0 (0.0)
0.7 (0.4) 0.3 (0.1) 0.1 (0.1) 0.1 (0.0) 0.1 (0.1) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)
Mar. 0.7 (1.8) 1.2 (1.4) 2.2 (2.1) 2.7 (5.2) 1.6 (1.5) 2.2 (1.6) 2.6 (3.4) 1.3 (0.7) 4.6 (3.0) −2.2 (4.7) −0.9 (1.7) −0.2 (0.9) −0.1 (0.5) −0.5 (1.2) −0.2 (0.5) −0.1 (0.3) −0.1 (0.5) 0.0 (0.0)
Apr.
May
0.1 (0.1) 0.1 (0.0) 0.4 (0.1) 0.5 (0.4) 0.2 (0.2) 0.4 (0.1) 0.5 (0.2) 0.3 (0.0) 0.8 (0.3)
0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)
−2.8 (15.2) −3.7 (12.5) −2.9 (9.5) −1.9 (5.9) −3.1 (8.0) −2.0 (4.8) −1.7 (3.9) −2.5 (6.8) −0.3 (0.6)
−1.8 (25.7) −2.0 (25.2) −3.1 (23.5) −3.3 (23.0) −3.9 (23.4) −4.1 (21.2) −4.6 (19.9) −3.3 (21.8) −4.1 (9.1)
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29 Table 3 Changes in the length of fruiting period (the difference in days between the first square to last effective square) for the three different sowing times. Locations
Earlier sowing
Normal sowing
Later sowing
Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
19.5 22.0 18.5 15.8 18.5 16.8 16.8 20.5 13.8
18.8 21.0 17.8 13.3 17.0 14.8 16.0 18.5 13.0
18.8 18.3 14.8 12.0 14.5 13.8 13.8 16.8 13.0
possible impact and the extent of adaptation needed. This study considered both changes in the mean climate and in climate variability. This implied greater T variation and that our management paradigms that are based on averages will have to shift in the future. 4.2. Consequences of changes in phenology Our results show that the start of key phenological events such as the emergence (1–9 d), first square (4–13 d), first flower (5–14 d) and first open boll (8–16 d) advanced across locations and GCMs (Fig. 3a). Based on experimental study, Reddy et al. (1996) found that the 1st square, 1st flower and 1st open boll were advanced by 1.8 d, 3.3 d and 7.3 d, respectively, per degree of warming. These represent an advancement of 2 d, 4 d and 9 d with a temperature increase of 1.2 ◦ C in line with our study. The extent of advancement is close to or within the range of our findings. Different cultivars, baseline Ts and management regimes may have contributed to difference in phenology development between the two studies. Positive impacts on changes in development were related to early emergence and a potential increase in the reproductive (fruiting) period as a result of earlier first square and delayed last effective square. The delay in the last effective square was a result of both a delay in the onset of first frost (≤2 ◦ C) event in the autumn and a more rapid increase in boll development. More rapid seed germination at all locations will lead to increased crop vigour and allows crops to be better established with a more flexible sowing window (Bange et al., 2008b; Braunack et al., 2012). The period from first square to last effective square is effective period when fruit is developed. This study showed that the period (Table 3) could be increased by 2–3 weeks, which would result in increased yields approximately 136 kg/ha lint for each additional week in delay in cutout and subsequent maturity (Bange and Milroy, 2004b). Changes in sowing times did not increase this period in all locations (Table 3). The earlier sowing increased the period 1–2 days while the later reduced the period 0–3 days. Adding to improved potential vigour early in growth and supporting better growth at the end of the season there were reductions in the number of cold shocks across locations in responding to the increase of Tmin (Table 2). Potential negative effects associated with changes in development may be associated with (1) reductions in vegetative growth period to support reproductive growth, (2) a loss of reproductive capacity due to reduced boll filling periods, and (3) an increase in fruit shedding arising from high T-stress days. For first square and first flower a more rapid advancement in all locations means that reproductive growth will occur earlier placing greater demands on overall growth. While Reddy et al. (1999) showed that vegetative growth could be increased with elevated T and CO2 concentrations, one consideration may be that there will be shorter time for vegetative growth to support high fruit loads on the plant. If less vegetative growth is produced without proper management
27
or cultivars to suit, this may cause crops to ‘cutout’ more quickly and reduce yield potential. Cutout is the cessation of fruit production when the demand on the resource supply by growing fruit increases to a point where none remains for the initiation and support of new fruiting sites (Hearn, 1969, 1972). Reduction in the time to ‘cutout’ will bring on earlier maturity, thus reducing yield (Bange and Milroy, 2004b). Yeates et al. (2010) investigated the effects of early season irrigation on vegetative growth and found that more vegetative growth was necessary to support high yielding cotton systems that used transgenic cotton with high and early fruit loads. There was also evidence that boll periods will be shorter as demonstrated by the periods from first flower to first open boll and from last effective flower to last harvestable boll. This is a similar finding to Reddy et al. (1999) who also observed shorter cotton boll development. As a consequence, fibre properties were influenced and boll size was smaller. Despite potential increases in the fruiting periods which may result in more fruit (bolls) the impact on yield may be tempered by the reductions in boll size. Importantly, in terms of negative effects there were substantial increases in the incidences of heat stress days in some months and locations; a doubling in some cases. Increase in Tmax has led to increase in the no. of heat stress events across locations and GS months (except for May) with warmer months and places increased more (Table 2). Both the degree of increase in Tmax (Fig. 2) and current Tmax contributed to this spatial pattern. The chances for exceeding Tmax threshold at Emerald is higher than that of Hillston where smaller increase in Tmax was projected and lower current Tmax exists compared with the former location. Reddy et al. (1997) suggested that one of the main losses in reproductive capacity in cotton with elevated Ts associated with climate change was increased fruit abscission. It was suggested that necessary practices such as changes in sowing time and heat tolerant cultivars will be needed to overcome this constraint. Wilson et al. (2003), however, showed that with a longer season length with crop growth maintained, loss of fruit can be compensated to some degree with new growth, although increased season length coupled with warmer T will require greater resources (water and nutrition) to achieve similar or higher yield outcomes (Constable and Bange, 2006). No account for changes in crop development associated with fruit abscission was investigated in this study and would need to be considered in future analyses. A higher temperature may increase the frequency, abundance and development rate of some insects [i.e. Silver leaf whitefly, Gregg and Wilson (2008)]. There will be more abundant winter and summer weed hosts of cotton pests and more disease (e.g. Alternaria macrospore) outbreaks in a wetter and warmer environment. Changes in these dynamics and their potential impact on cotton development have not been considered in this analysis.
5. Conclusion This analysis highlights the challenges associated with temperature with future climate change for future cotton production in Australia. The results show that there will be less impact of cold temperatures and a longer growing season that can be beneficial for cotton production. On the other hand, there will be more incidences of heat stress impacting growth, and more rapid crop development towards crop maturity that may limit the opportunities associated with increases in growing season length without adjustments in management. Results from this study will help to inform more complex analyses of climates changes impacts on cotton production in Australia and assist into development of adaptation strategies especially with the use of high yielding transgenic cotton with early high fruit loads in Australian high yielding (>2000 kg lint/ha) irrigated systems.
28
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
Even though this work has a focus on agro-ecosystem, the concepts, methodologies, and presentation framework adopted in this study can be applied to other disciplines such as plant/forest ecology, ecosystem science, entomology, and biodiversity. Advanced phenology arising from a warmer environment requests management options need to be put in place to avoid the potential negative impacts and/or to take the opportunity of global warming. Acknowledgements We thank Drs McGregor and Nguyen, CSIRO Marine and Atmospheric Research for providing us the high resolution climate change projections from four GCMs and Mr Chris Bull for putting extracted data onto the tape. Thanks also go to Dr Semenov, Rothamsted Research, UK for providing us the weather generator, LARS-WG, and Mr Peter Devoil for letting us access long times series of historical climate datasets. This project is financially supported by the Australian Cotton Research and Development Corporation with project number: UTS1301. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.ecolmodel.2014.04.018. References Bange, M.P., Milroy, S.P., 2004a. Impact of short term exposure to cold temperatures on early development of cotton (Gossypium hirsutum L.). Aust. J. Agric. Res. 55, 655–664. Bange, M.P., Milroy, S.P., 2004b. Growth and dry matter partitioning of diverse cotton genotypes. Field Crops Res. 87, 73–87. Bange, M.P., Caton, S.J., Milroy, S.P., 2008a. Managing yields of high fruit retention in transgenic cotton (Gossypium hirsutum L.) using sowing date. Aust. J. Agric. Res. 59, 733–741. Bange, M.P., McRae, D., Roth, G., 2008b. Cotton. In: Stokes, C.J., Howden, S.M. (Eds.), An Overview of Climate Change Adaptation in Australian Primary Industries – Impacts, Options and Priorities. CSIRO, Canberra, Australia, pp. 71–93. Bange, M.P., Constable, G.A., Gordon, S.G., Naylor, G.R.S., Van der Sluijs, M.H.J., 2009. A Guide to Improving Australian Cotton Fibre Quality. CSIRO and the Cotton Catchment Communities Coop. Research Centre, Narrabri, NSW, Australia, pp. 30–42. Bange, M.P., Constable, G.A., Johnston, D.A., Kelly, D., 2010. A method to estimate the effects of temperature on cotton micronaire. J. Cott. Sci. 14, 164–172. Barnston, A.G., Mason, S.J., Goddard, L., Dewitt, D.G., Zebiak, S., 2003. Multimodel ensembling in seasonal climate forecasting at IRI. Bull. Am. Meteorol. Soc. 84, 1783–1796. Bauer, P.J., Frederick, J.R., Bradow, J.M., Sadler, E.J., Evans, D.E., 2000. Canopy photosynthesis and fiber properties of normal- and late-planted cotton. Agron. J. 92, 518–523. Braunack, M.V., Bange, M.P., Johnston, D.B., 2012. Can planting date and cultivar selection improve resource use efficiency of cotton systems? Field Crops Res. 137, 1–11. Burke, J.J., Mahan, J.R., Hatfield, J.L., 1988. Crop-specific thermal kinetic windows in relation to wheat and cotton biomass production. Agron. J. 80, 553–556. Burke, J.J., Velten, J., Oliver, M.J., 2004. In vitro analysis of cotton pollen germination. Agron. J. 96, 359–368. Christiansen, M.N., 1967. Periods of sensitivity to chilling in germination cotton. Plant Physiol. 42, 431–433. Christiansen, M.N., Thomas, R.O., 1969. Season-long effects of chilling treatments applied to germinating cottonseed. Crop Sci. 9, 672–673. Conaty, W.C., Burke, J.J., Mahan, J.R., Neilsen, J.E., Sutton, B.G., 2012. Determining the optimum plant temperature of cotton physiology and yield to improve plantbased irrigation scheduling. Crop Sci. 52, 1828–1836. Constable, G.A., 1991. Mapping the production and survival of fruit on field grown cotton. Agron. J. 83, 374–378. Constable, G.A., Bange, M.P., 2006. What is cotton’s sustainable yield potential. The Australian Cottongrower 26 (7), 6–10. Constable, G.A., Shaw, A.J., 1988. Temperature Requirements for Cotton. Agfact P5.3.5. Division of Plant Industries, New South Wales Department of Agriculture. Constable, G.A., Harris, N.V., Paull, R.E., 1976. The effect of planting date on the yield and some fibre properties of cotton in the Namoi Valley. Aust. J. Exp. Agr. Anim. Husb. 16, 265–271.
CSIRO and Bureau of Meteorology, 2007. Climate Change in Australia, report. Australian Greenhouse Office, http://www. Technical climatechangeinaustralia.gov.au, 148 pp. Dong, H., Li, W., Tang, W., Li, Z., Zhang, D., Niu, Y., 2006. Yield, quality and leaf senescence of cotton grown at varying planting dates and plant densities in the Yellow River Valley of China. Field Crops Res. 98, 106–115. Gipson, J.R., 1974. Effect of temperature and methyl parathion on vegetative development and fruiting of the cotton plant. Agron. J. 66, 337–341. Gipson, J.R., Joham, H.E., 1968. Influence of night temperature on growth and development of cotton (Gossypium hirisutum L.). II. Fiber properties. Agron. J. 60, 296–298. Gipson, J.R., Joham, H.E., 1969. Influence of night temperature on growth and development of cotton (Gossypium hirisutum L.). III. Fiber elongation. Crop Sci. 9, 127–129. Gregg, P.C., Wilson, L.J., 2008. The changing climate for entomology. In: Proceedings of the 14th Australian Cotton Conference, August 12–14, 2008, Australian Cotton Growers Research Association, Broadbeach, Queensland. Hagedorn, R., Doblas-Reyes, F.J., Palmer, T.N., 2005. The rationale behind the success of multi-model ensembles in seasonal forecasting. I. Basic concept. Tellus A 57, 219–233. Hearn, A.B., 1969. The growth and performance of cotton in a desert environment. II. Dry matter production. J. Agric. Sci., Camb. 73, 75–86. Hearn, A.B., 1972. The growth and performance of rain-grown cotton in a tropical upland environment. II. The relationship between yield and growth. J. Agric. Sci., Camb. 79, 137–145. Hearn, A.B., Constable, G.A., 1984. Cotton. In: Goldsworthy, P.R., Fisher, N.M. (Eds.), The Physiology of Tropical Field Crops. John Wiley and Sons, Ltd., Chichester, UK, pp. 495–527. Hodges, H.F., Reddy, K.R., McKinion, J.M., Reddy, V.R., 1993. Temperature effects on cotton. Mississippi Agricultural and Forestry Experiment Station Bull 990. Mississippi State, pp. 1–15. Kakani, V.G., Reddy, K.R., Koti, S., Wallace, T.P., Prasad, P.V.V., Reddy, V.R., Zhao, D., 2005. Differences in in vitro pollen germination and pollen tube growth of cotton cultivars in response to high temperature. Ann. Bot. 96, 59–67. Karl, T.R., Nicholls, N., Gregory, J., 1997. The coming climate. Scientific American May, 55–59. Luo, Q., 2011. Temperature thresholds and crop production: a review. Clim. Change 109 (3), 583–598. Luo, Q., Yu, Q., 2012. Developing higher resolution climate change scenarios for agricultural risk assessment: progress, challenges and prospects. Int. J. Biometeorol. 56 (4), 557–568. Luo, Q., Bellotti, W.D., Hayman, P., Williams, M., Devoil, P., 2010. Effects of changes in climatic variability on agricultural production. Clim. Res. 42, 111–117. Mauney, J.R., 1986. Vegetative growth and development of fruiting sites. In: Mauney, J.R., Stewart, J.McD (Eds.), Cotton Physiology. Cotton Foundation, Memphis, TN, pp. 11–28. McGregor, J.L., Dix, M.R., 2008. An updated description of the conformal-cubic atmospheric model. In: Hamilton, K., Ohfuchi, W. (Eds.), High Resolution Simulation of the Atmosphere and Ocean. Springer, New York, pp. 51–76. McMahon, J., Low, A., 1972. Growing degree days as a measure of temperature effects on cotton. Cott. Gr. Rev. 49, 39–49. Nakicenovic, N., Swart, R., 2000. Emissions Scenarios: Special Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, UK, pp. 570. Pettigrew, W.T., 2008. The effect of higher temperature on cotton lint yield production and fiber quality. Crop Sci. 48, 278–285. Reddy, K.R., Hodges, H.F., McKinion, J.M., Wall, G.W., 1992a. Temperature effects on Pima cotton growth and development. Agron. J. 84, 237–243. Reddy, K.R., Hodges, H.F., Reddy, V.R., 1992b. Temperature effects on cotton fruit retention. Agron. J. 84, 26–30. Reddy, K.R., Hodges, H.F., McCarty, W.H., McKinion, J.M., 1996. Weather and Cotton Growth: Present and Future. Mississippi State University, Starkville, USA. Reddy, K.R., Hodges, H.F., McKinion, J.M., 1997. A comparison of scenarios for the effect of global climate change on cotton growth and yield. Aust. J. Plant Physiol. 24, 707–713. Reddy, K.R., Davidonis, G.H., Johnson, A.S., Vinyard, B.T., 1999. Temperature regime and carbon dioxide enrichment alter cotton boll development and fibre properties. Agron. J. 91, 851–858. Reddy, K.R., Hodges, H.F., Kimball, B.A., 2000. Crop ecosystem responses to climatic change: cotton. In: Reddy, K.R., Hodges, H.F. (Eds.), Climate Change and Global Crop Productivity. CAB International, New York, NY, pp. 161–187. Reddy, K.R., Prasad, P.V.V., Kakani, V.G., 2005. Crop responses to elevated carbon dioxide and interactions with temperature. Cott. J. Crop Improv. 13, 157–191. Robertson, B., Bednarz, C., Burmester, C., 2007. Growth and development – first 60 days. cotton physiology today. Newsletter of the Cotton Physiology Education Program – National Cotton Council, 13 (2). Snider, J.L., Oosterhuis, D.M., Skulman, B.W., Kawakami, E.M., 2009. Heat stressinduced limitations to reproductive success in Gossypium hirsutum. Physiol. Plant. 137, 125–138. Stiller, W.N., Reid, P.E., Constable, G.A., 2004. Maturity and leaf shape as traits influencing cotton cultivar adaptation to dryland conditions. Agron. J. 96, 656–664. Weigel, A.P., Liniger, M.A., Appenzeller, C., 2008. Can multi-model combination really enhance the prediction skill of probabilistic ensemble forecasts? Q. J. R. Meteorol. Soc. 134, 241–260.
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29 Wilson, L.J., Sadras, V.O., Heimoana, S.C., Gibb, D., 2003. How to succeed by doing nothing: cotton compensation after simulated early season pest damage. Crop Sci. 73, 2125–2134. Yeates, S.J., Roberts, J., Richards, D., 2010. High insect protection of GM Bt cotton changes crop morphology and response of water compared to non Bt cotton. In: Proceedings of the 15th ASA Conference, Australian
29
Society of Agronomy, Lincoln, New Zealand http://www.regional.org.au/ au/asa/2010/crop-production/physiology-breeding/7046 yeatess.htm Yoon, S.T., Hoogenboom, G., Flitcroft, I., Bannayan, M., 2009. Growth and development of cotton (Gossypium hirsutum L.) in response to CO2 enrichment under two different temperature regimes. Environ. Exp. Bot. 67, 178–187.
Contents lists available at ScienceDirect
Ecological Modelling journal homepage: www.elsevier.com/locate/ecolmodel
Cotton crop phenology in a new temperature regime Qunying Luo a,∗ , Michael Bange b , Loretta Clancy b a b
University of Technology, Sydney, PO Box 123, Broadway, NSW 2007, Australia CSIRO Division of Plant Industry, Locked Bag 59, Narrabri, NSW 2390, Australia
a r t i c l e
i n f o
Article history: Received 1 February 2014 Received in revised form 15 April 2014 Accepted 16 April 2014 Available online 13 May 2014 Keywords: Temperature increase Cotton phenology Heat stress Cold shock
a b s t r a c t The daily outputs of the CSIRO Conformal Cubic Atmospheric Model driven by four general circulation models (GCMs) were used by a stochastic weather generator, LARS-WG, to construct four local climate change scenarios (CCSs) at nine key cotton production areas in eastern Australia. These CCSs were then linked to daily temperature-driven models of cotton phenology to examine the magnitude of the effects of increased temperature on the initiation and duration of key crop phenophases and on the occurrence of heat stress and cold shocks during the growing season. The results show that when using 1st Oct. as sowing time (1) the timing of emergence, 1st square, 1st flower and 1st open boll advanced 1–9, 4–13, 5–14 and 8–16 days, respectively, for the period centred on 2030 compared to baseline; (2) when crops were planted 10 days earlier, the emergence stage occurred earlier in most of the locations while other phenological events changed slightly (∼1 day) in comparison with 1st Oct. sowing; when crops were planted 10 days later, all these events were generally delayed (∼1.5 days) in comparison with 1st Oct. sowing; (3) the timing of the last effective square, last effective flower and last harvestable boll were delayed 7–12, 6–9 and 3–9 days, respectively, across locations (except St George) and GCMs; (4) the fruit development period increased up to 2–3 weeks; (5) the number of hot days increased across all locations and growing season (GS) months except May with the warmer months (Dec., Jan. and Feb.) and locations increased more; and (6) the number of cold shocks decreased or remained the same across locations and GS months except Jan. and Feb. with cold months and places decreased more. The results show that there will be less impact of cold temperatures on earlier growth and potentially a longer growing season that can improve crop yield. However, there will be more incidences of hot days impacting growth, and more rapid crop development in late phenological stages (especially during boll filling) that may limit the opportunities associated with increases in growing season length without adjustments in management. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Ambient temperature (T) is one of the major environment variables affecting the growth, development and yields of crops, especially the rate of development (Luo, 2011). Crops have basic requirements for T to complete a specific phenophase or the whole life cycle. On the other hand, extremely high and low temperatures (Ts) can have detrimental effects on crop growth, development and yield, particularly at critical phenophases. While cotton is morphologically indeterminate, the rate of many phenological processes such as germination, floral initiation and development of fruiting bodies is controlled by T (Hearn and Constable, 1984). Average daily T also plays an important role in determining the earliest date of sowing, defining season length which can both influence yield potential and cotton fibre quality (Bange et al., 2008a; Bauer et al.,
∗ Corresponding author. Tel.: +61 0410 186 352; fax: +61 2 9514 4079. E-mail address: [email protected] (Q. Luo). http://dx.doi.org/10.1016/j.ecolmodel.2014.04.018 0304-3800/© 2014 Elsevier B.V. All rights reserved.
2000; Dong et al., 2006), and determining where cotton can be produced sustainably. Generally the longer the growing season (GS) the greater the potential for higher lint yields (Bange and Milroy, 2004a; Stiller et al., 2004). In Australian cotton production systems, T requirements for the development of cotton are described by the accumulation of degree days calibrated with a base T of 12 ◦ C (Constable and Shaw, 1988). Other cotton systems elsewhere (i.e. America) have used a base T closer to 15 ◦ C (Robertson et al., 2007). Throughout most Australian production systems, minimum temperature (Tmin ) (≤11 ◦ C) experienced early in the cotton season can cause delays in development, reduction in growth, and sometimes chilling injury (Constable et al., 1976; Bange and Milroy, 2004b). As the season progresses maximum Ts (>35 ◦ C, heat stress) are commonplace throughout cotton production regions and may adversely affect growth and development thus affecting water use, yield and fibre quality (Hodges et al., 1993). At the end of a cotton season cooler or cold T will influence the timing of crop maturity and impact on effectiveness of chemical harvest aids, both again directly affecting yield and quality. Cotton has an optimal thermal
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
23
Fig. 1. Cotton production locations in Eastern Australia used in this study.
kinetic window of 23–32 ◦ C in which metabolic activity is most efficient (Burke et al., 1988; Conaty et al., 2012). High T affects reproductive processes of cotton thus affecting yield (Reddy et al., 2000, 2005). Effects may be attributed to impacts on pollen viability (Burke et al., 2004), fertilisation efficiency (Snider et al., 2009), boll size and seed number (Pettigrew, 2008), and fruit shedding (Hodges et al., 1993; Reddy et al., 1992b). Kakani et al. (2005) reported that instantaneous air T above 32 ◦ C reduces cotton pollen viability, and T above 29 ◦ C reduces pollen tube elongation. Reddy et al. (2005) demonstrated that the maximum growth rate per boll occurred at 25–26 ◦ C and decreased at higher Ts while boll harvest index was highest at 28 ◦ C, declining thereafter and reaching zero boll harvest index at 33–34 ◦ C. Boll size was largest when Ts less than 20 ◦ C. Percent boll set, boll number, boll filling period, rate of boll growth, boll size and yield all decreased when mean T reached 30 ◦ C (Reddy et al., 2005). T of 40/32 ◦ C led to zero boll yield (Reddy et al., 1992a,b). Heat stress can result in boll cavitation and reduces cotton lint yield (Bange et al., 2008b). From a fibre quality perspective sustained increases in T will reduce fibre length (Gipson and Joham, 1969) and increase micronaire (Gipson and Joham, 1968; Bange et al., 2010), and sometimes negatively impacting final fibre quality. Cotton exposed to low Ts takes longer to develop, and accumulates biomass at a slower rate (Gipson, 1974; Mauney, 1986). Cold T responses (below 10 ◦ C) in early post-emergent seedlings can permanently arrest growth and development (Christiansen, 1967; Christiansen and Thomas, 1969). Bange and Milroy (2004b) also found that negative effects on development existed when plants were exposed to at least 10 nights at 10 ◦ C, or for 5 nights at 5 ◦ C on post-emergent cotton. In Australia, to account for early season effects of cold T on cotton development, a ‘cold shock’ effect is applied to degree days (DDs) accumulation. A ‘cold shock’ is defined
as an event when the daily minimum T falls to 11 ◦ C or less, which assumed to cause chilling injury and thus delays development. Late in the season cold Ts (≤2 ◦ C) can signal the end of the GS. These low T may force bolls to open affecting lint fibre quality (colour and maturity), and will severely impede the efficacy of chemical harvest aids (optimal T ∼18 ◦ C) to remove leaf from the plant prior to harvest (Bange et al., 2009). The optimum sowing time aims to reduce the incidences of ‘cold shocks’, while in-season management aims to ensure that fruit has adequate time to mature and that harvest aids are applied prior to the onset of cold T. Increase in T will change crop phenology, including the start and duration of phenophases and the probability and degree of extreme climate events such as heat stress and cold shocks. These changes may have significant implications for cotton lint yield and fibre quality. Hence this work aims to understand local climate change in Australian cotton production regions, quantify its impacts on cotton crop phenology and on the occurrence of heat stress and cold shocks in cotton GS. This work will support more complex studies and analyses into climate change on cotton production and assist in identifying effective adaptation options to deal with the risk and to take the advantage of climate change. 2. Materials and methods 2.1. Study locations This study focused on nine major cotton production areas in Queensland (Emerald, Dalby, St George and Goondiwindi) and New South Wales (Moree, Bourke, Narrabri, Warren and Hillston) in Australia (Fig. 1). These cotton production areas also represent different growing environments: with Emerald, Bourke and St George being classified as being hot (annual mean T: 20.21–22.61 ◦ C);
24
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
Dalby, Goondiwindi, Moree, Narrabri and Warren as being mild (annual mean T: 18.21–19.91 ◦ C); and Hillston as cool (annual mean T of 17.58 ◦ C). Classification of these climates for cotton production was based on the analysis of McMahon and Low (1972) using growing DD. These different environments have resulted in adoption of different crop management practices and cultivars. 2.2. Local climate change and climate change scenarios Dynamic downscaling is one of the major downscaling approaches for translating coarser spatial resolution climate change information to finer scales (Luo & Yu, 2012). In this study, the outputs of the CSIRO Conformal Cubic Atmospheric Model (CCAM), a dynamic downscaling approach, for baseline (1980–1999) and future period (2020–2039) were used by a stochastic weather generator, LARS-WG, to derive local T change including changes in the mean (i.e., Tmax , Tmin ) and in variability (i.e., mean T variability). For the performance of this weather generator, readers are directed to Online Resource 1. The CCAM model was driven by four GCMs specifically GFDL, CSIRO Mark 3.5, MPI, and MIROC, under the Special Report on Emission Scenarios (SRES, Nakicenovic and Swart, 2000) A2 scenario. The rationale of using the outputs of CCAM driven by four GCMs is that multi-model ensemble is intrinsically more useful and accurate than sets of projections produced with any single model (Barnston et al., 2003; Hagedorn et al., 2005; Weigel et al., 2008). More information on CCAM is provided by McGregor and Dix (2008) and Luo et al. (2010). The local T change information derived earlier for each location were then reapplied to the LARSWG along with the historical daily T data to construct long time series (100y) temperature scenarios including baseline and future scenarios for impact assessment. The period 1980–1999 is standard baseline period. We chose 2020–2039 as the future period as farmers are more interested in near timeframe climate change rather than many (i.e. 40) years later. 2.3. Cotton crop phenology and occurrence of extreme temperatures Cotton sowing in the Australian industry occurs around 1st Oct. (Bange et al., 2008a). In this analysis, we chose three sowing times: 21st Sep., 1st Oct. and 11th Oct. We then defined the GS as the period from these sowing dates until 21st May, 31st May and 10th Jun., respectively. For the mild and cool locations these sowing dates generally represent a range of sowing dates currently used. Cotton crop phenology stages considered in this study include the date of emergence, first square (flower bud), first flower, first open boll, last effective square (LES), last effective flower (LEF), and last harvestable boll (LHB). The DD targets corresponding to the first four stages are set to 80, 505, 777 and 1527 DD, respectively (Constable and Shaw, 1988). The growth targets used have been calibrated with hot days occurring during development. The last harvestable boll is associated with the occurrence of the first frost, which is defined as the Tmin of ≤2 ◦ C. The date of the last effective square, flower, and boll can be used to determine target ‘cutout’ dates that assist in ceasing production of new fruiting sites to allow crops (and all bolls) to be mature in time for harvest. The DD differences between the date of LHB and LEF (LHB period) and between the LEF and LES (square period) are set to 750 and 430 DD, respectively (Constable, 1991). The DD were derived using a base T of 12 ◦ C (Constable and Shaw, 1988): DD =
[(Tmax − 12) + (Tmin − 12)] 2
(1)
where Tmax and Tmin are daily maximum and minimum Ts, respectively. When Tmin < 12 ◦ C, Tmin = 12 (or (Tmin − 12) ≥ 0.0).
As mentioned in Section 1, to account for cold T on early season cotton development a ‘cold shock’ effect is applied to DD accumulation for predicting the first square and first flower. A ‘cold shock’ is defined as an event when daily Tmin falls to 11 ◦ C or less, and extends the DD accumulation by 5.2 DD for each cold shock event (Hearn and Constable, 1984). Extreme Ts considered in this study include the occurrence of heat stress and cold shock on a monthly basis within the GS (1 Oct.–31 May). A heat stress event occurs when daily Tmax is ≥35 ◦ C. To quantify the effects of climate change on crop phenology, the on-line DD Calculator and Last Effective Flower tool (http://cottassist.com.au), developed by CSIRO, were modified to be used with baseline (1980–1999) and future period (2020–2039) T scenarios for the study locations. A programme was edited to quantify the occurrence of cold shocks and heat stress on a monthly basis. Changes in cotton phenology and in the occurrence of heat stress and cold shocks were presented in multi-model ensemble means, which is the average derived from the four driving GCMs. The analysis also assumes that there are no effects of elevated CO2 on the rates of cotton development. Reddy et al. (1999) found that there is no direct effects of enhanced CO2 on phenology, however, a more recent studies by Yoon et al. (2009) at very high levels of CO2 (800 mol l−1 ) found that there was more rapid development to flowering at moderate daytime Ts (25 ◦ C day/15 ◦ C night). It should be noted that the occurrence of late phenological events and of cold shocks and heat stress is independent of sowing times. 3. Results 3.1. Temperature change Fig. 2 shows the multi-model ensemble mean changes between the periods centred on 2030 and 1989 in Tmax , Tmin , and Tmean for GS and for each location. Tmax (0.9–1.1 ◦ C), Tmin (1.1–1.3 ◦ C) and Tmean (1–1.2 ◦ C) all increased across locations with Tmin increasing the most (Fig. 2). The largest change in average T occurred in Dalby, associated with a considerable shift in both Tmax and Tmin . The smallest change in Tmean was at Hillston associated with the lowest changes in both Tmax and Tmin across locations. The largest (climate) model-to-model variation (error bar) associated with Tmax was found at St George while the least was found at Emerald and Hillston. The increase in Tmax is greater than 0.6 ◦ C for all models and locations. For Tmin , the least model-to-model difference was found at Emerald and the others were fairly similar. The increase in Tmin was greater than 1 ◦ C for all models and locations. 3.2. Changes in crop phenology 3.2.1. Normal sowing Fig. 3 shows the multi-model ensemble mean changes in key phenological events for the period centred on 2030 in comparison with the baseline period 1980–1999. T increase accelerated the phenological development of all phases across all locations (Fig. 3). Emergence, first square, first flower and first open boll individually were advanced 1–9, 2–8, 1–2 and 1–5 days, respectively (Fig. 3a). The largest changes were projected in the period from sowing to emergence, especially for the locations further south. In general the smallest effect across all phases was associated with the most northern growing region of Emerald while the largest change for all phases was associated with the most southern growing areas of Warren and Hillston. This translated into the smallest overall cumulative change for Emerald (mean change of 8 days) and greatest change for Hillston and Warren (mean change 16 days). For the end of season, the timing of the phenological stages of the LES, LEF and LHB were delayed 0.3–3, 1.8–6.8 and 2.5–9 days,
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
25
a
Emerald Dalby St George Goondiwindi Moree Bourke Narrabri Warren Hillston 0.6
0.8
1.0
1.2
1.4
Changes in Growing Season Maximum Temperature (oC)
b
Emerald Dalby St George Goondiwindi Moree Bourke Narrabri Warren Hillston
0.6
0.8
1.0
1.2
1.4
Changes in Growing Season Minimum Temperature (oC)
Emerald
c
Dalby St George Goondiwindi Moree Fig. 3. Cumulative contribution of each phenophases to cotton crop phenology in 2030: (a) earlier phenological stages associated with normal sowing (sown on the 1st of Oct.) in comparison with baseline and (b) later phenological stages.
Bourke Narrabri Warren Hillston 0.6
0.8
1.0
1.2
1.4
Changes in Growing Season Mean Temperature (oC) Fig. 2. Changes in growing season (1st Oct.–31st May) average temperatures for 2030: (a) GS Tmax , (b) GS Tmin and (c) GS Tmean . Horizontal line is the error bar, which is calculated as the standard deviation divided by root mean square of the sample size. The empty circle is the multi-model ensemble mean.
respectively, across all locations except at Warren where the LEF was unaffected and at St George where the LHB advanced slightly (Fig. 3b). The greatest variability in delays was associated with LHB (at T ≤ 2 ◦ C) with the largest changes projected for Warren (mean change 9 d), Dalby (mean change 7 d) and Narrabri (mean change
6 d). The smallest changes for LHB were for St George (mean change −0.3 d) and Bourke (mean change 2.5 d). Across all locations, LES and LEF were also later as the periods between LHB to LEF, and LEF to LES were shorter. Overall, this meant that the smallest overall delays in season end were for Bourke (mean change 6.5 d), Hillston (mean change 7.3 d) and St George (mean change 7.5 d). The projected longest delays in season end were for Dalby (mean change 11.8 d) followed by Moree (mean change 10 d). 3.2.2. Earlier sowing With earlier sowing (21st Sept.), emergence advanced by 1–3 d or was unchanged across locations except for St George where a 3 d delay was found. First square was advanced by 1–4 d in northern cotton production areas (with warmer climate) but was delayed ∼1 d in southern areas (with colder climate). In general there was
26
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
Table 1 Changes (in days) in early phenological events of an earlier (21 Sep.) and later sowing (11 Oct.) in comparison with 1 Oct. sowing in a changing climate. Locations
Emergence
1st square
1st flower
1st open boll
Earlier sowing Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
−1.5 −1.0 −1.5 −3.0 −2.3 0.0 3.0 −1.5 0.0
0.8 0.0 0.8 0.5 0.8 −2.0 −3.8 −0.5 −0.8
0.5 0.3 0.5 −0.5 0.3 0.5 −0.8 0.5 0.8
−0.5 −0.3 0.0 0.5 0.0 0.5 1.0 0.5 −0.3
Later sowing Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
−0.3 0.0 2.8 2.3 4.0 0.5 2.3 3.3 0.5
0.3 2.8 0.3 −1.0 −1.5 0.5 0.0 −1.5 −0.5
1.3 0.8 −0.5 −0.5 0.3 −0.5 −0.5 −0.5 0.5
−1.5 −0.8 0.0 −0.3 −0.3 1.0 0.5 0.0 −0.3
decreased or unchanged across locations and GS months except Jan. and Feb. where minor increase occurred in some of the locations especially NSW cotton production areas. Generally speaking, relatively larger decrease in the no. of cold shocks was found at colder/southern locations (i.e. Hillston & Warren) and vice versa for the warmer places (i.e. Emerald) in most of the months. The minor increase in the number of cold shocks in Jan. and Feb. at some of the locations especially NSW cotton production areas reflected the effects of changed T variability on this parameter as increased T variability can lead to both increase of hot days and cold shocks. Both spatial and temporal variability exist associated with these two extreme Ts. 4. Discussion 4.1. Local temperature change
little effect on the 1st open boll as advancement or delay was no more than 1 d (Table 1) 3.2.3. Later sowing With later sowing (11th Oct.) in comparison with 1st Oct. sowing, emergence delayed 1–4 d except at Hillston where a 0.3 d advancement was found. First square was advanced by 1–2 d in warmer locations while delayed 0.3–3 d in colder locations. First flower and 1st open boll were advanced or delayed in the range of 0–1.5 d depending on locations (Table 1). 3.3. Occurrence of extreme temperatures Table 2 shows the frequency of heat stress days (>35 ◦ C) and cold shocks (≤11 ◦ C) within GS months for the nine cotton production areas under investigation. In 2030, the number of heat stress days increased across all locations and across GS months except May where no change was found. Larger increase occurred in warmer months (Nov.–Mar.) and warmer locations (i.e. Emerald, Goondiwindi and St George). On the contrary, the number of cold shocks
Tmax , Tmean and Tmin increased 0.9–1.1 ◦ C, 1–1.2 ◦ C and 1.1–1.3 ◦ C, respectively (Fig. 2) across locations and GCMs for the period centred on 2030 under A2 emission scenario, which represents a medium–high emission scenario. Overall, increase in Tmin is larger than that of Tmax . This finding is in line with earlier climate change studies conducted in other countries (i.e. Karl et al., 1997). Even though all four GCMs projected an increase in Ts, model-tomodel differences exist with larger difference found at St George and smaller at Hillston and Emerald associated with Tmax . CSIRO and BOM (2007) reported an increase of 1–1.5 ◦ C (median projection) for the inner part of Australia under medium–high emission scenario. Our results are very close to the results reported by CSIRO and BOM (2007). Spatial variation exists in the increase of T indicating different management options and/or efforts are needed to tackle the effects of climate change in different cotton production regions. This analysis considered uncertainties due to climate modelto-model difference. The local climate change was derived from a dynamical downscaling approach: one of the mainstream downscaling approaches. The limitation of this analysis is that only one emission scenario (A2) was considered due to the large computing resources needed to undertake this type of downscaling approach. Nonetheless it represents a relatively high emission scenario with corresponding higher increase in T compared with other low emission scenarios. Therefore impact analysis and adaptation evaluation based this emission scenario should capture the upper range of the
Table 2 Changes in the frequency of heat stress events (>35 ◦ C) and cold shocks (≤11 ◦ C) within GS months (Oct.–May inclusive). Numbers in brackets are the frequency of extreme Ts under baseline climate conditions. Oct. Heat stress Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
0.5 (0.5) 0.7 (0.5) 0.8 (0.4) 1.8 (1.3) 0.8 (0.5) 1.4 (0.7) 1.8 (1.1) 1.5 (0.4) 2.9 (2.9)
Cold shocks Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
−5.1 (19.4) −6.2 (18.2) −5.3 (13.9) −3.9 (7.9) −3.1 (11.0) −2.6 (6.8) −2.7 (5.9) −3.0 (8.8) −0.6 (1.0)
Nov. 0.9 (1.8) 1.8 (2.2) 1.9 (2.5) 3.1 (5.9) 2.3 (2.4) 2.2 (2.6) 4.8 (4.5) 1.9 (1.3) 6.2 (6.9) −4.3 (8.4) −2.5 (4.9) −1.8 (3.1) −0.5 (1.0) −1.3 (2.9) −0.5 (1.2) −0.4 (0.8) −0.7 (1.6) 0.0 (0.0)
Dec. 1.6 (6.8) 3.7 (7.2) 5.0 (7.7) 4.4 (14.9) 4.6 (6.0) 5.7 (6.6) 5.3 (11.8) 3.0 (3.1) 5.8 (13.9) −1.7 (2.1) −0.6 (0.9) −0.4 (0.6) −0.1 (0.1) −0.3 (0.5) −0.1 (0.1) 0.0 (0.0) −0.1 (0.2) 0.0 (0.0)
Jan.
Feb.
3.3 (10.1) 2.1 (12.4) 2.6 (11.6) 0.0 (19.3) 3.4 (9.3) 3.4 (10.0) 0.5 (15.6) 2.9 (3.4) 5.4 (12.6)
2.0 (7.1) 2.6 (6.7) 2.9 (6.9) 2.3 (12.6) 3.9 (5.6) 3.7 (5.7) 2.4 (9.4) 1.8 (2.0) 3.1 (6.8)
0.3 (0.5) 0.1 (0.1) 0.0 (0.1) 0.1 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.1) 0.0 (0.0)
0.7 (0.4) 0.3 (0.1) 0.1 (0.1) 0.1 (0.0) 0.1 (0.1) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)
Mar. 0.7 (1.8) 1.2 (1.4) 2.2 (2.1) 2.7 (5.2) 1.6 (1.5) 2.2 (1.6) 2.6 (3.4) 1.3 (0.7) 4.6 (3.0) −2.2 (4.7) −0.9 (1.7) −0.2 (0.9) −0.1 (0.5) −0.5 (1.2) −0.2 (0.5) −0.1 (0.3) −0.1 (0.5) 0.0 (0.0)
Apr.
May
0.1 (0.1) 0.1 (0.0) 0.4 (0.1) 0.5 (0.4) 0.2 (0.2) 0.4 (0.1) 0.5 (0.2) 0.3 (0.0) 0.8 (0.3)
0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0) 0.0 (0.0)
−2.8 (15.2) −3.7 (12.5) −2.9 (9.5) −1.9 (5.9) −3.1 (8.0) −2.0 (4.8) −1.7 (3.9) −2.5 (6.8) −0.3 (0.6)
−1.8 (25.7) −2.0 (25.2) −3.1 (23.5) −3.3 (23.0) −3.9 (23.4) −4.1 (21.2) −4.6 (19.9) −3.3 (21.8) −4.1 (9.1)
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29 Table 3 Changes in the length of fruiting period (the difference in days between the first square to last effective square) for the three different sowing times. Locations
Earlier sowing
Normal sowing
Later sowing
Hillston Warren Narrabri Bourke Moree Goondiwindi St George Dalby Emerald
19.5 22.0 18.5 15.8 18.5 16.8 16.8 20.5 13.8
18.8 21.0 17.8 13.3 17.0 14.8 16.0 18.5 13.0
18.8 18.3 14.8 12.0 14.5 13.8 13.8 16.8 13.0
possible impact and the extent of adaptation needed. This study considered both changes in the mean climate and in climate variability. This implied greater T variation and that our management paradigms that are based on averages will have to shift in the future. 4.2. Consequences of changes in phenology Our results show that the start of key phenological events such as the emergence (1–9 d), first square (4–13 d), first flower (5–14 d) and first open boll (8–16 d) advanced across locations and GCMs (Fig. 3a). Based on experimental study, Reddy et al. (1996) found that the 1st square, 1st flower and 1st open boll were advanced by 1.8 d, 3.3 d and 7.3 d, respectively, per degree of warming. These represent an advancement of 2 d, 4 d and 9 d with a temperature increase of 1.2 ◦ C in line with our study. The extent of advancement is close to or within the range of our findings. Different cultivars, baseline Ts and management regimes may have contributed to difference in phenology development between the two studies. Positive impacts on changes in development were related to early emergence and a potential increase in the reproductive (fruiting) period as a result of earlier first square and delayed last effective square. The delay in the last effective square was a result of both a delay in the onset of first frost (≤2 ◦ C) event in the autumn and a more rapid increase in boll development. More rapid seed germination at all locations will lead to increased crop vigour and allows crops to be better established with a more flexible sowing window (Bange et al., 2008b; Braunack et al., 2012). The period from first square to last effective square is effective period when fruit is developed. This study showed that the period (Table 3) could be increased by 2–3 weeks, which would result in increased yields approximately 136 kg/ha lint for each additional week in delay in cutout and subsequent maturity (Bange and Milroy, 2004b). Changes in sowing times did not increase this period in all locations (Table 3). The earlier sowing increased the period 1–2 days while the later reduced the period 0–3 days. Adding to improved potential vigour early in growth and supporting better growth at the end of the season there were reductions in the number of cold shocks across locations in responding to the increase of Tmin (Table 2). Potential negative effects associated with changes in development may be associated with (1) reductions in vegetative growth period to support reproductive growth, (2) a loss of reproductive capacity due to reduced boll filling periods, and (3) an increase in fruit shedding arising from high T-stress days. For first square and first flower a more rapid advancement in all locations means that reproductive growth will occur earlier placing greater demands on overall growth. While Reddy et al. (1999) showed that vegetative growth could be increased with elevated T and CO2 concentrations, one consideration may be that there will be shorter time for vegetative growth to support high fruit loads on the plant. If less vegetative growth is produced without proper management
27
or cultivars to suit, this may cause crops to ‘cutout’ more quickly and reduce yield potential. Cutout is the cessation of fruit production when the demand on the resource supply by growing fruit increases to a point where none remains for the initiation and support of new fruiting sites (Hearn, 1969, 1972). Reduction in the time to ‘cutout’ will bring on earlier maturity, thus reducing yield (Bange and Milroy, 2004b). Yeates et al. (2010) investigated the effects of early season irrigation on vegetative growth and found that more vegetative growth was necessary to support high yielding cotton systems that used transgenic cotton with high and early fruit loads. There was also evidence that boll periods will be shorter as demonstrated by the periods from first flower to first open boll and from last effective flower to last harvestable boll. This is a similar finding to Reddy et al. (1999) who also observed shorter cotton boll development. As a consequence, fibre properties were influenced and boll size was smaller. Despite potential increases in the fruiting periods which may result in more fruit (bolls) the impact on yield may be tempered by the reductions in boll size. Importantly, in terms of negative effects there were substantial increases in the incidences of heat stress days in some months and locations; a doubling in some cases. Increase in Tmax has led to increase in the no. of heat stress events across locations and GS months (except for May) with warmer months and places increased more (Table 2). Both the degree of increase in Tmax (Fig. 2) and current Tmax contributed to this spatial pattern. The chances for exceeding Tmax threshold at Emerald is higher than that of Hillston where smaller increase in Tmax was projected and lower current Tmax exists compared with the former location. Reddy et al. (1997) suggested that one of the main losses in reproductive capacity in cotton with elevated Ts associated with climate change was increased fruit abscission. It was suggested that necessary practices such as changes in sowing time and heat tolerant cultivars will be needed to overcome this constraint. Wilson et al. (2003), however, showed that with a longer season length with crop growth maintained, loss of fruit can be compensated to some degree with new growth, although increased season length coupled with warmer T will require greater resources (water and nutrition) to achieve similar or higher yield outcomes (Constable and Bange, 2006). No account for changes in crop development associated with fruit abscission was investigated in this study and would need to be considered in future analyses. A higher temperature may increase the frequency, abundance and development rate of some insects [i.e. Silver leaf whitefly, Gregg and Wilson (2008)]. There will be more abundant winter and summer weed hosts of cotton pests and more disease (e.g. Alternaria macrospore) outbreaks in a wetter and warmer environment. Changes in these dynamics and their potential impact on cotton development have not been considered in this analysis.
5. Conclusion This analysis highlights the challenges associated with temperature with future climate change for future cotton production in Australia. The results show that there will be less impact of cold temperatures and a longer growing season that can be beneficial for cotton production. On the other hand, there will be more incidences of heat stress impacting growth, and more rapid crop development towards crop maturity that may limit the opportunities associated with increases in growing season length without adjustments in management. Results from this study will help to inform more complex analyses of climates changes impacts on cotton production in Australia and assist into development of adaptation strategies especially with the use of high yielding transgenic cotton with early high fruit loads in Australian high yielding (>2000 kg lint/ha) irrigated systems.
28
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29
Even though this work has a focus on agro-ecosystem, the concepts, methodologies, and presentation framework adopted in this study can be applied to other disciplines such as plant/forest ecology, ecosystem science, entomology, and biodiversity. Advanced phenology arising from a warmer environment requests management options need to be put in place to avoid the potential negative impacts and/or to take the opportunity of global warming. Acknowledgements We thank Drs McGregor and Nguyen, CSIRO Marine and Atmospheric Research for providing us the high resolution climate change projections from four GCMs and Mr Chris Bull for putting extracted data onto the tape. Thanks also go to Dr Semenov, Rothamsted Research, UK for providing us the weather generator, LARS-WG, and Mr Peter Devoil for letting us access long times series of historical climate datasets. This project is financially supported by the Australian Cotton Research and Development Corporation with project number: UTS1301. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.ecolmodel.2014.04.018. References Bange, M.P., Milroy, S.P., 2004a. Impact of short term exposure to cold temperatures on early development of cotton (Gossypium hirsutum L.). Aust. J. Agric. Res. 55, 655–664. Bange, M.P., Milroy, S.P., 2004b. Growth and dry matter partitioning of diverse cotton genotypes. Field Crops Res. 87, 73–87. Bange, M.P., Caton, S.J., Milroy, S.P., 2008a. Managing yields of high fruit retention in transgenic cotton (Gossypium hirsutum L.) using sowing date. Aust. J. Agric. Res. 59, 733–741. Bange, M.P., McRae, D., Roth, G., 2008b. Cotton. In: Stokes, C.J., Howden, S.M. (Eds.), An Overview of Climate Change Adaptation in Australian Primary Industries – Impacts, Options and Priorities. CSIRO, Canberra, Australia, pp. 71–93. Bange, M.P., Constable, G.A., Gordon, S.G., Naylor, G.R.S., Van der Sluijs, M.H.J., 2009. A Guide to Improving Australian Cotton Fibre Quality. CSIRO and the Cotton Catchment Communities Coop. Research Centre, Narrabri, NSW, Australia, pp. 30–42. Bange, M.P., Constable, G.A., Johnston, D.A., Kelly, D., 2010. A method to estimate the effects of temperature on cotton micronaire. J. Cott. Sci. 14, 164–172. Barnston, A.G., Mason, S.J., Goddard, L., Dewitt, D.G., Zebiak, S., 2003. Multimodel ensembling in seasonal climate forecasting at IRI. Bull. Am. Meteorol. Soc. 84, 1783–1796. Bauer, P.J., Frederick, J.R., Bradow, J.M., Sadler, E.J., Evans, D.E., 2000. Canopy photosynthesis and fiber properties of normal- and late-planted cotton. Agron. J. 92, 518–523. Braunack, M.V., Bange, M.P., Johnston, D.B., 2012. Can planting date and cultivar selection improve resource use efficiency of cotton systems? Field Crops Res. 137, 1–11. Burke, J.J., Mahan, J.R., Hatfield, J.L., 1988. Crop-specific thermal kinetic windows in relation to wheat and cotton biomass production. Agron. J. 80, 553–556. Burke, J.J., Velten, J., Oliver, M.J., 2004. In vitro analysis of cotton pollen germination. Agron. J. 96, 359–368. Christiansen, M.N., 1967. Periods of sensitivity to chilling in germination cotton. Plant Physiol. 42, 431–433. Christiansen, M.N., Thomas, R.O., 1969. Season-long effects of chilling treatments applied to germinating cottonseed. Crop Sci. 9, 672–673. Conaty, W.C., Burke, J.J., Mahan, J.R., Neilsen, J.E., Sutton, B.G., 2012. Determining the optimum plant temperature of cotton physiology and yield to improve plantbased irrigation scheduling. Crop Sci. 52, 1828–1836. Constable, G.A., 1991. Mapping the production and survival of fruit on field grown cotton. Agron. J. 83, 374–378. Constable, G.A., Bange, M.P., 2006. What is cotton’s sustainable yield potential. The Australian Cottongrower 26 (7), 6–10. Constable, G.A., Shaw, A.J., 1988. Temperature Requirements for Cotton. Agfact P5.3.5. Division of Plant Industries, New South Wales Department of Agriculture. Constable, G.A., Harris, N.V., Paull, R.E., 1976. The effect of planting date on the yield and some fibre properties of cotton in the Namoi Valley. Aust. J. Exp. Agr. Anim. Husb. 16, 265–271.
CSIRO and Bureau of Meteorology, 2007. Climate Change in Australia, report. Australian Greenhouse Office, http://www. Technical climatechangeinaustralia.gov.au, 148 pp. Dong, H., Li, W., Tang, W., Li, Z., Zhang, D., Niu, Y., 2006. Yield, quality and leaf senescence of cotton grown at varying planting dates and plant densities in the Yellow River Valley of China. Field Crops Res. 98, 106–115. Gipson, J.R., 1974. Effect of temperature and methyl parathion on vegetative development and fruiting of the cotton plant. Agron. J. 66, 337–341. Gipson, J.R., Joham, H.E., 1968. Influence of night temperature on growth and development of cotton (Gossypium hirisutum L.). II. Fiber properties. Agron. J. 60, 296–298. Gipson, J.R., Joham, H.E., 1969. Influence of night temperature on growth and development of cotton (Gossypium hirisutum L.). III. Fiber elongation. Crop Sci. 9, 127–129. Gregg, P.C., Wilson, L.J., 2008. The changing climate for entomology. In: Proceedings of the 14th Australian Cotton Conference, August 12–14, 2008, Australian Cotton Growers Research Association, Broadbeach, Queensland. Hagedorn, R., Doblas-Reyes, F.J., Palmer, T.N., 2005. The rationale behind the success of multi-model ensembles in seasonal forecasting. I. Basic concept. Tellus A 57, 219–233. Hearn, A.B., 1969. The growth and performance of cotton in a desert environment. II. Dry matter production. J. Agric. Sci., Camb. 73, 75–86. Hearn, A.B., 1972. The growth and performance of rain-grown cotton in a tropical upland environment. II. The relationship between yield and growth. J. Agric. Sci., Camb. 79, 137–145. Hearn, A.B., Constable, G.A., 1984. Cotton. In: Goldsworthy, P.R., Fisher, N.M. (Eds.), The Physiology of Tropical Field Crops. John Wiley and Sons, Ltd., Chichester, UK, pp. 495–527. Hodges, H.F., Reddy, K.R., McKinion, J.M., Reddy, V.R., 1993. Temperature effects on cotton. Mississippi Agricultural and Forestry Experiment Station Bull 990. Mississippi State, pp. 1–15. Kakani, V.G., Reddy, K.R., Koti, S., Wallace, T.P., Prasad, P.V.V., Reddy, V.R., Zhao, D., 2005. Differences in in vitro pollen germination and pollen tube growth of cotton cultivars in response to high temperature. Ann. Bot. 96, 59–67. Karl, T.R., Nicholls, N., Gregory, J., 1997. The coming climate. Scientific American May, 55–59. Luo, Q., 2011. Temperature thresholds and crop production: a review. Clim. Change 109 (3), 583–598. Luo, Q., Yu, Q., 2012. Developing higher resolution climate change scenarios for agricultural risk assessment: progress, challenges and prospects. Int. J. Biometeorol. 56 (4), 557–568. Luo, Q., Bellotti, W.D., Hayman, P., Williams, M., Devoil, P., 2010. Effects of changes in climatic variability on agricultural production. Clim. Res. 42, 111–117. Mauney, J.R., 1986. Vegetative growth and development of fruiting sites. In: Mauney, J.R., Stewart, J.McD (Eds.), Cotton Physiology. Cotton Foundation, Memphis, TN, pp. 11–28. McGregor, J.L., Dix, M.R., 2008. An updated description of the conformal-cubic atmospheric model. In: Hamilton, K., Ohfuchi, W. (Eds.), High Resolution Simulation of the Atmosphere and Ocean. Springer, New York, pp. 51–76. McMahon, J., Low, A., 1972. Growing degree days as a measure of temperature effects on cotton. Cott. Gr. Rev. 49, 39–49. Nakicenovic, N., Swart, R., 2000. Emissions Scenarios: Special Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, UK, pp. 570. Pettigrew, W.T., 2008. The effect of higher temperature on cotton lint yield production and fiber quality. Crop Sci. 48, 278–285. Reddy, K.R., Hodges, H.F., McKinion, J.M., Wall, G.W., 1992a. Temperature effects on Pima cotton growth and development. Agron. J. 84, 237–243. Reddy, K.R., Hodges, H.F., Reddy, V.R., 1992b. Temperature effects on cotton fruit retention. Agron. J. 84, 26–30. Reddy, K.R., Hodges, H.F., McCarty, W.H., McKinion, J.M., 1996. Weather and Cotton Growth: Present and Future. Mississippi State University, Starkville, USA. Reddy, K.R., Hodges, H.F., McKinion, J.M., 1997. A comparison of scenarios for the effect of global climate change on cotton growth and yield. Aust. J. Plant Physiol. 24, 707–713. Reddy, K.R., Davidonis, G.H., Johnson, A.S., Vinyard, B.T., 1999. Temperature regime and carbon dioxide enrichment alter cotton boll development and fibre properties. Agron. J. 91, 851–858. Reddy, K.R., Hodges, H.F., Kimball, B.A., 2000. Crop ecosystem responses to climatic change: cotton. In: Reddy, K.R., Hodges, H.F. (Eds.), Climate Change and Global Crop Productivity. CAB International, New York, NY, pp. 161–187. Reddy, K.R., Prasad, P.V.V., Kakani, V.G., 2005. Crop responses to elevated carbon dioxide and interactions with temperature. Cott. J. Crop Improv. 13, 157–191. Robertson, B., Bednarz, C., Burmester, C., 2007. Growth and development – first 60 days. cotton physiology today. Newsletter of the Cotton Physiology Education Program – National Cotton Council, 13 (2). Snider, J.L., Oosterhuis, D.M., Skulman, B.W., Kawakami, E.M., 2009. Heat stressinduced limitations to reproductive success in Gossypium hirsutum. Physiol. Plant. 137, 125–138. Stiller, W.N., Reid, P.E., Constable, G.A., 2004. Maturity and leaf shape as traits influencing cotton cultivar adaptation to dryland conditions. Agron. J. 96, 656–664. Weigel, A.P., Liniger, M.A., Appenzeller, C., 2008. Can multi-model combination really enhance the prediction skill of probabilistic ensemble forecasts? Q. J. R. Meteorol. Soc. 134, 241–260.
Q. Luo et al. / Ecological Modelling 285 (2014) 22–29 Wilson, L.J., Sadras, V.O., Heimoana, S.C., Gibb, D., 2003. How to succeed by doing nothing: cotton compensation after simulated early season pest damage. Crop Sci. 73, 2125–2134. Yeates, S.J., Roberts, J., Richards, D., 2010. High insect protection of GM Bt cotton changes crop morphology and response of water compared to non Bt cotton. In: Proceedings of the 15th ASA Conference, Australian
29
Society of Agronomy, Lincoln, New Zealand http://www.regional.org.au/ au/asa/2010/crop-production/physiology-breeding/7046 yeatess.htm Yoon, S.T., Hoogenboom, G., Flitcroft, I., Bannayan, M., 2009. Growth and development of cotton (Gossypium hirsutum L.) in response to CO2 enrichment under two different temperature regimes. Environ. Exp. Bot. 67, 178–187.