Báo cáo khoa học: "A comparison of daily representations of canopy conductance based on two conditional timeaveraging methods and the dependence of daily conductance on environmental factors"

Chia sẻ: Nguyễn Minh Thắng | Ngày: | Loại File: PDF | Số trang:19

Thêm vào BST

Báo xấu

76
lượt xem 2
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

Tuyển tập các báo cáo nghiên cứu về lâm nghiệp được đăng trên tạp chí lâm nghiệp quốc tế đề tài: "A comparison of daily representations of canopy conductance based on two conditional timeaveraging methods and the dependence of daily conductance on environmental factors...

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: Báo cáo khoa học: "A comparison of daily representations of canopy conductance based on two conditional timeaveraging methods and the dependence of daily conductance on environmental factors"

Original article comparison of daily representations of canopy A conductance based on two conditional time- averaging methods and the dependence of daily conductance on environmental factors Ram Oren Nathan Phillips NC 27708-0328, USA Nicholas School of the Environment, Duke University, Durham, (Received 15 January 1997; accepted 23 September 1997) Abstract - In hydrological models which incorporate vegetated surfaces, non-steady state responses in stem sap flow to diurnal evaporative demand can lead to unreasonable values of com- puted canopy conductance, which corrupt diurnal courses and daily averages. Conductance computations based on daily averaged constituent variables are a potential method for circum- venting this problem. However, comparisons between these two averaging methods are lacking. In this study, both methods for computing daily canopy conductance were compared in a pine for- est. A simplification of the Penman-Monteith equation under conditions of high aerodynamic cou- pling was used to calculate instantaneous canopy conductance. Large variation between the two methods was observed due to biases introduced under conditions of low sap flow or vapor pres- sure deficit. Two conditional averaging schemes were developed to exclude data which were strongly affected by such conditions, and as a result of the conditional averaging, a tighter rela- tionship between these two averaging schemes was found. We calculated daily representations of canopy conductance for an entire growing season in a 15-year-old Pinus taeda stand. Despite clear declines in conductance between rain events in direct response to soil water depletion, and large seasonal dynamics in canopy leaf area, canopy conductance remained generally uniform until low late season temperatures. (© Inra/Elsevier, Paris.) Pinus taeda / canopy conductance / sap flux / soil water balance Résumé - Comparaison des estimations de conductance de couvert journalière basées sur deux méthodes de moyennes temporelles conditionnelles, et effet des facteurs de l’envi- ronnement. Dans les modèles hydrologiques qui prennent en compte les surfaces végétales, les estimations de la conductance du couvert pour la vapeur d’eau faites à partir des mesures de flux de sève et de la demande climatique peuvent conduire à des valeurs instantanées et des * Correspondence and reprints (1) 919 613 8032; fax: (1) 919 684 8741; e-mail: ramoren@duke.edu Tel:
moyennes journalières erronées, à cause d’un régime de flux hydrique non conservatif. Un cal- cul de la conductance basé sur les moyennes journalières de certaines variables est une méthode pour résoudre ce problème. Toutefois, la comparaison entre les deux méthodes n’avait pas encore été effectuée. Ce travail compare ces deux méthodes de calcul de la conductance de couvert pour une forêt de pins. L’équation de Penman-Monteith a été simplifiée, en supposant un fort cou- plage entre le couvert et l’atmosphère, et utilisée ici pour calculer les valeurs instantanées de conductance de couvert. Un écart important entre les deux méthodes a été mis en évidence, à cause des biais apparaissant en conditions de flux de sève ou de déficit de saturation de l’air faibles. Deux procédures de moyennage conditionnel ont été développées pour exclure ce type de données, et une relation étroite entre les deux méthodes a été trouvée. La conductance de couvert journa- lière a été ainsi calculée sur une saison de végétation complète dans un peuplement de Pinus taeda âgé de 15 ans. En dépit de la diminution de conductance entre les épisodes de pluie, condui- sant à des déficits hydriques dans le sol, et de variations importantes d’indice foliaire au cours de la saison, la conductance du couvert est restée uniforme jusqu’aux basses températures de la fin de la saison. (© Inra/Elsevier, Paris.) Pinus taeda / conductance de couvert / flux de sève / bilan hydrique 1. INTRODUCTION [ 15, 46]. Other modeling approaches have used midday conductance information (e.g. [39]) or fixed values (see Shuttle- The development of proper methods worth [37]) as representative of daily con- for averaging canopy stomatal conduc- ductance. However, it is necessary to tance, both spatially and temporally, examine how well such values generalize remains a subject of active research. In diurnal patterns of conductance. In this temporal representations of canopy con- study we evaluate differences in conduc- ductance, the daily time scale is particu- tance values calculated based on mean larly important. Many climate or hydro- daily conditions versus those based on logical models that involve canopy averages of diurnal values. conductance use or predict information on this time scale (e.g. [3, 6, 39, 45, 47]); The use of stem sapflow measurements moreover, many climatological data are inconjunction with meteorological mea- available only at the daily scale [47]. How- surements of evaporative demand has pro- ever, an appropriate daily representation of vided a method in which continuous esti- canopy conductance lacks consensus (e.g mates of canopy conductance may be compare the approaches in Tattori et al. made [12, 13]. The advent of this tech- [41]; Couralt et al. [6]; Fennessey and nology has alleviated the constraint of Vogel [7]; Kustas et al. [23]), in spite of extrapolating conductances over time improvements in the ability to obtain con- based on steady state environmental rela- ductance information diurnally [13, 30]. tionships, or using single points within a Extrapolation from maximum diurnal val- day as representative of daily conductance. ues of leaf or canopy conductance to aver- However, the ability to continuously mea- aged daily conductance estimates based flux and environmental driving sure stem on environmental relationships [16] has forces for use in conductance calculations yielded reasonable results (e.g. [34, 35]), has presented a new challenge: resolving but this approach depends on steady state diurnal non-steady state canopy and stem relationships between conductance and characteristics and their effects on calcu- environmental variables, which may not lations of instantaneous and daily average be applicable within diurnal time scales conductance.
flux rates? Having addressed these ques- Leaf and canopy conductances have tions, the proposed study suggests a gen- been demonstrated to exhibit time lags eral, robust computational and conditional with respect to changes in environmental averaging scheme for generating daily influences [46]. Furthermore, due to canopy conductances in experiments effects of hydraulic capacitance and resis- where diurnal conductance values are tance, sap fluxes measured in stems of available. Further, the dependence of such trees lag behind canopy transpiration [13, a daily canopy conductance on environ- 32, 36]. Unless the time constants for these mental factors such as rain and soil mois- processes can be determined precisely and ture depletion is investigated. generally over a range of environmental conditions (e.g. water stress) the combined presence of both of these effects may lead to errors in instantaneous computations of 2. MATERIALS AND METHODS canopy conductance. Such instantaneous errors may then be propagated to daily The study was conducted in Duke Forest, estimates. North Carolina, USA, in a managed, 15-year- old Loblolly pine stand. More details on the An alternative method for estimating site are given in Phillips et al. [31]. Leaf area conductance which potentially daily index (L), derived from allometric relation- avoids the problems presented by averag- ships between leaf area and sapwood area, cou- pled with a foliage dynamics model [18], ing instantaneous diurnal values is a cal- ranged from 2.01to 3.27 during this period. culation based on daily averages of the Measurements needed to compute canopy con- constituent variables. In principle, the total ductance for 20 individuals were initiated on 4 daily transpiration could be related to the April 1995, and continued until 28 December total daily driving force, without the com- 1995. All measurements were made at 30-s plications arising from improper diurnal intervals and averaged and recorded every 30 matching of the two. The objective of this min with a data logger (Delta-T Devices, UK). study was to investigate how such a ) s xylem -2 m 20 H (g1 was estimated Sap flow method for calculation of average daily with constant-heat probes as described in canopy conductance (hereafter referred to Granier [11, 12]. Sap flux densities of indi- vidual trees were scaled to the plot level by differed as ’daily’ conductance, or C,day G ) multiplying average sap flux densities by the from one which used daily averages of amount of xylem area per unit ground area in diurnal values of conductance (hereafter the stand, corrected for a reduced flux rate referred to as ’mean diurnal’ conductance, found in inner xylem [29]. C,diu G ). or Air temperature and relative humidity were The specific questions addressed by measured at the bottom of the upper third of the forest canopy with a Vaisala HMP 35C this study were: if an inverse form of the temperature and humidity probe (Vaisala, Penman-Monteith equation is used to cal- Helsinki, Finland). Air vapor pressure deficit culate conductance, how will averaged (D) was calculated from air temperature and daily thermodynamic variables, based on relative humidity according to Goff and Gratch average daily temperature, differ from [10]. Photosynthetically active radiation (PAR) those obtained as the average of diurnal, was measured above the canopy with a spher- temperature-dependent measurements? ical quantum sensor (LI-193SA, Licor, Inc., Lincoln, Nebraska, USA). Rainfall was mea- How will nighttime uptake of water (either sured to the nearest 0.1 mm in a forest clearing from normal recharge or rain recharge) ca 100 m away from the site with a Texas Elec- differentially affect results of the two aver- tronics Tipping Bucket Rain Gage (Texas Elec- aging techniques? Will the difference tronics, Dallas, Texas, USA). However, mid- between a daily and diurnal conductance way through the season, this instrument failed. be a function of tree size or absolute daily Therefore, subsequent rain data were obtained
from a National Weather Service station located ca 5 km from the site. 2.1. Diurnal versus daily where U is windspeed (m s at height z (11.5 ) -1 conductance calculations m), z is roughness length (set as 1/10 of 0 canopy height, or 1 m), d is zero plane dis- placement (taken as 2/3 of canopy height, or 6.7 All conductances were calculated according m). Over the 6-d period, G computed accord- C to Monteith and Unsworth [26] as ing to equation (1) averaged 4.5 mm s (SE -1 = 0.2 mm s only 2.4 % of G which aver- ), -1 A , aged 187 mm s(SE =1 mm sAt one point -1 ). -1 within the 6-d period, G reached, at maxi- C mum, 10.9 % of G Thus, the assumption that . A G » G made in equation (1) appeared to CA where yis the psychrometric constant, λ is the be satisfied. latent heat of vaporization of water, is the density C liquid of The different averaging procedures for specific heat of air, p is the equation (1) compared in this study are sum- water, T is temperature, E is evaporation rate, C marized in table I and are discussed in further and D is vapor pressure deficit of the canopy detail in the following sections. air. This simplification of the Penman-Mon- teith equation is based on the assumption that the pine forest was well coupled aerodynami- cally [17]; thus D approximated the total driv- 2.2. Diurnal calculation of canopy ing force for transpiration. We checked the conductance, G C,diu assumption of strong coupling in this forest by selecting a six-d period from 23-28 May 1995 for diurnal comparisons of aerodynamic con- Equation (1) was solved at every 30-m step, ductance (G with total canopy conductance using E D and temperature at that time to ) A , C (G Aerodynamic conductance was com- ). C calculate the appropriate thermodynamic con- puted according to Thom and Oliver [43] as stants. Data were averaged over the period each
may be subject to large relative errors. These 0 μmol ms (hereafter -2-1 when PAR day > introduced under conditions of low errors are referred to daylight hours). as absolute daily transpiration (figure 1). 2.3. Daily calculation of canopy 2.4. Comparison of daily versus conductance, G C,day diurnal conductance values Equation (1) was solved using averages of A direct comparison between diurnal and allquantities over the day. Values of the ther- daily averaged conductance was made to indi- modynamic variables were based on the aver- cate the most general differences between the two methods over all conditions of E and D. age temperature over daylight hours. Daily C vapor pressure deficit In order to assess the influence of extremely (D was also obtained ) D by averaging D over daylight hours, under the low D values and their possible cause of assumption that this is the time period in which extremely high apparent conductance values, a D has an effect on transpiration and, therefore, conditional averaging scheme (hereafter referred to as the diurnal conditional average) uptake. Transpiration (E however, was ), C summed over 24 h but divided by daylight was applied to the diurnal values so as to hours only, because: 1), this accounted for all exclude all D values < 0.1 kPa. The 0.1-kPa water uptake driven by D over the day; and 2) cutoff was selected after observing unreason- it provided a consistent averaging period with ably high and low conductance values obtained that used for D Starting and ending points forD < 0.1 kPa. (the reasons for which will be . for days were taken as from 0500 to 0500 hours stated in the results section). This conditional (steady state conditions), after concluding that average (denoted by brackets) can be expressed using a 0000 to 0000 hours integration period in symbols as
(= 0.85) between 0- and a 60-min lag a for D Thus could not justify . C E x we applying a constant time lag to the entire data set to correct for non-steady state behavior. Conductance calculations were therefore based on D and E with zero C lag. 3.2. Comparison of daily versus diurnalthermodynamic constants The approximation of temperature- and N is the number of diurnal observations. dependent thermodynamic variables as The number of points, n, for each daytime period in which I 1 was recorded to assess constants (e.g. as defined at standard tem- i = the effect of sample size on the calculation of perature) in calculations of conductance daily conductance values. A secondary condi- can lead to significant errors if significant tion was applied to exclude days for which n temperature variation occurs. We calcu- was too small to provide reasonable statistical lated a 12 % relative difference in a com- representation of the daytime conductance bined conductance coefficient (table II) (hereafter referred to as the daily conditional between 0 and 35 °C due to temperature average). The choice of this n will be discussed in the results section. dependence of the conductance coeffi- cient. Thus, including temperature depen- dence into the thermodynamic variables 3. RESULTS involved in the conductance calculation is necessary under conditions of signifi- cant temperature variability. For the pur- 3.1. The potential of cross-correlation pose of calculating daily average canopy analysis to rectify non-steady conductance, this temperature dependence state behavior leads to question as to whether the ther- a modynamic constants can be averaged Cross-correlation analysis over the from diurnal values, or calculated from whole data set showed no difference in r average daily temperature, before being
averaged conductance, both when this con- used in equation (1). The answer to this ductance was obtained by averaging diur- question depends upon whether the tem- nal values (over light hours) or when using perature dependence of the thermody- the averaged constituent variables in a namic variables can be approximated as daily scale calculation (figure 2a). In addi- linear over the appropriate temperature range. Relationships between γ,λ and p tion, the difference between daily con- ductance based on diurnal values and that with temperature (C has very weak tem- p based on daily averages showed a large perature dependence [28]) are highly lin- divergence at low D (figure 2b). ear over the temperature range of physio- logical importance (defined for this study the range -5 to 35 °C; table II). Thus, as conclude that a combined conductance we 3.4. Effect of diurnal conditional coefficient can be used as a function of averaging of conductance values mean daytime temperature, for the pur- on the comparison between pose of calculating average daily conduc- diurnal and daily conductance according to equation (1). tance After examination of portions of our data set in which low diurnal vapor pres- 3.3. Pre-conditionally averaged sure deficit values occurred, both on mom- comparison between diurnal ings and evenings of clear days and and daily conductance throughout overcast diurnals (e.g., figure 3) a combination conditional average based Very large errors in computed diurnal on light level and a minimum D was conductance can result from conditions in defined. Conductance values in which I i which transpiration and/or vapor pressure 0 were excluded from averaging over = deficit have very low values [2]. In addi- the diurnal time series. The included and tion to lags between sap flux and vapor excluded data as a result of this condi- pressure deficit leading to magnified errors 3 tional averaging are illustrated in figure at low absolute values of vapor pressure for two typical diurnal time series. The deficit, slight biases in measurements of result of this averaging scheme on the rela- either vapor pressure deficit or sap flow tionship between computed daily and aver- may lead to diverging conductance val- age diurnal conductances is demonstrated ues when approaching a zero/zero ratio in figure 2c, d. The conditional averaging (specified error = ± 3 % in relative humid- led to a much reduced range of conduc- ity at > 90 % for the HMP35C transducer tance values both on a daily and average we used; also, sap flux measurements may diurnal basis (figure 2c). Similarly, the exhibit a small bias at low absolute val- difference between conductance values ues as a result of violation of assumptions obtained using the two methods at low D of zero flux at night [12]). Indeed, such decreased appreciably as compared with potential sources of error add to the moti- unconditionally averaged data (figure 2d vation for using daily averaged constituent compared with figure 2b). Still, the range variables in a daily conductance calcula- of values of conductance includes values tion. too large to be considered reasonable. In addition to very large values of conduc- We found that such unreasonably high tance, values approaching zero at low D low apparent instantaneous conduc- or may also be a result of small biases in tances, obtained under conditions where measurement where estimated E is rela- D < 0.1kPa, were influential enough to C tively smaller than D. lead to unreasonably high or low daily
3.5. Effect of daily conditional between daily and averaged diurnal ence averaging of conductances conductance calculations displays large vari- based on sample size ation. Thus we defined a secondary condi- tion which operated on whole days, using the criteria of a minimum sample size of n Figure 3 illustrates that during days in 12(or 6 h at our sampling rate) for inclu- which D stays low, a very small sampling = size may result for use in G Figure sion. The result of this second condition for 4 . C,diu demonstrates that when n < 12, the differ- acceptance is shown in figure 2e, f. The
ductance. Although the slope of daily cal- agreement between daily conductance and culated conductance (G to diurnal diurnally averaged conductance has been ) C,day further increased compared to figure 2c, d. averaged conductance (G was less ) C,diu This range in values is comparable with than one (slope = 0.96, P 0.001, inter- = typical diurnal values of conductance in 0.02), due to the inter- cept = 5.2e-4, P = pine forests from other studies (e.g. [9, 25]). was generally greater than cept termC,day G especially Still, several points (designated with trian- at low values. We inter- , C,diu G pret this to be caused by the inclusion of gle symbols) appear to lie outside of the general cluster of points. The specific con- water taken up in recharge at night in a ditions of those days were investigated in calculation of G which is not repre- C,day detail and found to result from rain sented in G thus leading to a relatively , C,diu were night. The effects of nighttime lower G We also expected this to be a events at . C,diu recharge in general and recharge due to rain function of the absolute magnitude of daily evaluated next. transpiration, as the ’proportion’ of events were recharge decreases (in the absence of rain) as daytime flux increases. Figure 5 shows 3.6. Effects of daytime versus that as night uptake becomes a significant nighttime flux on the difference fraction of total daily water uptake, GC,day between G and G increasingly exceeds G When night C,diu C,day . C,diu uptake exceeded ca 50 % of total uptake, The two conditional averaging schemes G decreased to ca 30 % of G At C,diu . C,day resulted in reasonable values of daily con- very low values of night uptake/total
most ratios of G clus- In addition to nighttime recharge due uptake, C,day :G C,diu to the lag between transpiration and stem tered around unity, but several points were sap flow, fast re-hydration of xylem capac- observed to reach higher values. Inspec- itance, which has been depleted by longer- tion of the original diurnal courses showed term drying periods, occurs owing to rain that these data came from days in which Such rain events may introduce events. daytime values of D occurred which were into G day or G calculations if C,diu C, errors only slightly greater than the threshold recharge is mistaken for evaporative- 0.1kPa, but nevertheless led to relatively demand driven sap flow. During daytime high calculated conductance values due to rain events, it is very difficult to separate moderate flux rates. For instance, on 19 sap flow driven by the environment from April 1995, three morning measurements sap flow due to recharge. However, rain of D (within light hours) that averaged only events which occur at night can lead to 0.35 kPa led to diurnal values of conduc- well-defined pulses of rain-induced tance that averaged 15.9 mm s while the , -1 recharge, permitting a quantification of rest of the daytime values (n 19) aver- = their effects on calculations of G ver- C,day 2.40 mm sThe effect of the . -1 affected aged only sus G because G is not C,diu, C,diu three large values was to inflate G to C,diu by nighttime rain recharge. Figure 6 pre- 4.25 mm swhile -1 was calculated as C,day G sents data taken from 9-d selected from 2.85 mm sclose to the 2.40 mm s of -1 , -1 the long-term data set in which nighttime the other 19 points. This illustrates that rain, and associated nighttime stem diurnal averages of conductance may be recharge, occurred. Although it cannot in highly sensitive to the choice of a threshold general be excluded that nighttime rain- induced recharge may contribute to water level of D used for diurnal conditional uptake during subsequent daylight hours, averaging. Thus, rather than finding a dis- the data shown in figure 6 were taken from tinct threshold at which D creates biases diurnal courses in which stable zero-flux in diurnal conductance averaging, a con- baselines were observed after night rain tinuous bias may be introduced into diurnal recharges, yet before transpiration started. conductance averages in which, as the con- It is apparent that at lower integrated day- ditional threshold is raised, a bias toward time stand transpiration, recharge of stor- lower averaged diurnal conductance may age depleted over the long term can be introduced. Additionally, changes in approach and even surpass the daytime the conditional threshold would lead to value. While nighttime rains will intro- changes in the average number of diurnal duce large errors into estimates, C,day G values accepted for averaging, potentially daytime rain, especially during days of affecting the statistical representation of low D and E will introduce such errors , C diurnal averages. Therefore a choice of a into both estimates of conductance. threshold D condition must balance 1) the effects of low point values of D and asso- ciated biases introduced as a result of both 3.7. Effect of tree size sensor error at low absolute rates, as well as nighttime-daytime flux on 2) the effects of non-steady state behavior of stem sap flux in relation to canopy D, If time taken for nighttime recharge against 3) the legitimate inclusion of high function of tree volume, it would were a conductance values occurring at low values not be possible to use in a calcula- C,day G ground-based of D and E and 4) the need for enough C tion an average, stand tran- data points within a daytime period to ade- spiration based on sap flow data from a range of tree sizes. Rather, each tree would quately characterize the daytime hours.
have to have its own conductance calcu- the seasonal of leaf-area on course lation, and the conductance computed weighted conductance, based on mean from all measured trees would then be daily conditions, is shown in figure 7. Also averaged. However, we did not find sig- shown are relevant environmental vari- nificant dependence of nighttime recharge ables. In the 266 d record, 8 d were lost on tree volume (P > 0.1; data not shown). as a result of missing or bad data associ- Thus, we conclude that in the forest stud- ated with sensor malfunction. Of the ied over the range of tree sizes monitored remaining 258 d, the defined threshold (dbh range 110-201 mm), conductance conditions led to the exclusion from con- = calculations may appropriately be per- ductance calculations of an additional 23 d. formed on the average flux of a sample of trees exhibiting a range of volumes. Fur- ther studies on a larger range of tree sizes 4. DISCUSSION would be necessary in order to generalize this finding. Use of stem sap flow measuring tech- offers a tool for investigating the niques issue of temporally varying fluxes and sur- 3.8. Resulting seasonal course face resistance, but this technique is com- of daily conductance plicated by time variation both at the canopy level and throughout the xylem. The result of the two-stage conditional Although the non-steady state behavior of averaging process employed in this study water fluxes through the soil-plant-atmo-
continuum has long been recog- flux. The discrepancy between results here sphere nized [20, 21], its effect on computations and in the previous study may involve the of canopy conductance when utilizing length and timing of the present study stem sap flux measurements has only (266 d through an entire growing season) versus that in the previous study (50 d in recently begun to be investigated. the early part of a growing season). Although detailed soil moisture informa- We have found that the simplest possi- tion was not available in the previous ble correction for non-steady state behav- study, based on precipitation input and ior of stem flow with respect to canopy early growing season conditions, soil flux - made from a lag analysis - was moisture was probably not limiting in that unable to satisfactorily identify a consistent study. However, it has been shown that time constant, thus, leading us to investi- the time constant for nighttime recharge gate further the utility and robustness of is an increasing function of soil moisture a daily conductance based on mean daily depletion in cases in which soil moisture is conditions. Our inability in this study to determine a consistent time lag which already relatively limiting [24, 31].Thus, it is probable that in the present study, could be used to rectify the non-steady long-term variations of soil moisture, state relationship between stem flux and canopy transpiration is in contrast with which included periods of significant earlier findings [32] in which a consistent moisture limitation, affected our ability to find a consistent time constant for stem time constant of about 45 min was esti- mated for stem flow responses to canopy response to canopy evaporative demand.
The earlier finding that nighttime daily means, G to a mean conduc- C,day on recharge increases as soil moisture weighted by above-canopy PAR tance becomes more depleted [31]presents (representing available energy). Figure 8 another reason for concern in attempts to demonstrates a good relationship between use averaged diurnal values of conduc- both G and G versus weighted C,diu C,day daily conductance according to Monteith tance in a daily conductance representa- tion. If nighttime recharge increases as a et al. [27], designated G Regres- . C,Monteith function of soil drying, then larger pro- sions for both G and G versus C,diu C,day portions of flux at night relative to flux G had slopes less than one (0.90 C,Monteith during the day (similar to that seen in fig- for G P < 0.0001; 0.88 for G , C,diu , C,day 0.0001), with slightly positive inter- ure 6), which would not be accounted for P < mm s for -1 in the diurnal-averaging approach, would G P < 0.0001; cepts (0.43 , C,diu 0.89 mm sfor G < 0.0001), indi- -1 lead to an increasing underestimation of ,P C,day conductance (figure 5). This problem cating that both G and G were C,diu C,day would be rectified when conductance cal- progressively lower than G asC,Monteith culations use mean daily conditions. daily conductance increases. A positive intercept and positive slope less than unity In this study, we have determined con- arises because of the effects of diurnal ditions in which a daily conductance based covariance of G with PAR. When G and on mean daily meteorological conditions PAR covary strongly, G is C,Monteith approaches that computed as the arith- weighed proportionally more relative to metic mean of diurnal conductance val- when G and PAR have little or negative ues (figure 2). We found two primary covariance. This pivots the slope down- conditions to be sufficiently high instan- wards and result in a positive intercept. taneous values of D (figure 3), and a long Scatter in the relationship between G C,diu enough period during the day in which and G was found in cases where C,Monteith such values occur (figure 4). Arithmetic 1) extremely high diurnal values of appar- means of diurnal surface conductance are ent conductance accompanying daytime often used as representations of daily con- recharge of long-term storage depletion ductance (e.g. [3]). However, it should be from rain were not adequately discounted noted that the arithmetic mean of diurnal by the weighted average procedure, while values may not be the best daily repre- such points were eliminated from the diur- sentation of conductance for use in pre- nal average because D < 0.1kPa, resulting dictions of total daily transpiration. The in a relatively greater calculated G C,Monteith mean of diurnal values may be greatly than G and 2) late afternoon rain , C,diu affected by a few unreasonable values events associated with very low D but sub- originating under conditions of low E C stantial stem recharge led to high apparent and D (figure 3; [2]). Furthermore, Mon- conductance values which were discounted teith et al. [27] argued that diurnal con- by the weighted averaging procedure since ductance values should be weighted by light was decreasing in the late afternoon, some measure of available energy for resulting in a relatively greater calculated evaporation, resulting in a daily weighted G than G C,diu . C,Monteith mean conductance. Variations of such weighting procedures have also been used For most applications involving daily by others [8, 38, 40]. surface conductance, it would be more Thus, in order to provide more utility desirable to have access to diurnal courses from study for hydrological models, our of conductance than daily means. How- we compared both arithmetic mean con- ever, when using stem sap flow tech- and conductance based ductance, , C,diu G niques, such diurnal courses are affected
suffices, a correction factor to G by non-steady state conditions, causing a C,day mean systematic underestimation in most con- may be approximated. For each day, the ditions. Therefore, there is utility in using ratio of G to G can be calcu- C,Monteith C,diu lated and multiplied by G This incor- daily means. Daily means of conductance . C,day porates the effect of diurnal weighting of cannot be directly adjusted to include the effects of weighted averaging such as was conductance by available energy and more carried out by Monteith et al. [27]. In closely reflects the weighted averaging of applications for which a weighted daily Monteith et al. [27], but suffers less from
the underestimation inherent in G We sites occupied with shallowly rooting . C,diu or found that the correction factor ranging species, soil moisture reserves are amply from null to two times available only if they are replenished fre- , C,day G was not or D, and was not dependent on quently by precipitation. In addition, for a C,day G significantly different from unity for the given species, higher L results in higher entire data set (1.006; S.E. 0.009). E [ 14], exhausting soil moisture in a given = layer faster. Thus, on a deep soil, a P. pinaster Ait. stand with low L showed a gradual decrease in canopy conductance to 4.1. Seasonal course of G C,day 50 % in late summer of values in early ca in relation to leaf area dynamics spring, as the soil dried slowly [24]. Tran- and soil moisture conditions spiration, however, was maintained high until mid-summer owing to increasing Seasonal patterns in G are strongly C,day potential evaporation. Only in late sum- affected by phenological and physical fac- mer did a combination of lower potential tors. In temperate forests, general pat- a evaporation and canopy conductance of increased canopy conductance tern finally reduce transpiration to 25 % of the accompanies the increase in leaf area. This high, spring values, a reduction similar to is very pronounced in deciduous forests, that reported for a P. radiata D. Don stand but is also clearly apparent in forests com- [42]. This seasonal pattern contrasts with posed of non-deterministic evergreen responses of a P. pinaster stand on shal- species with highly dynamic leaf produc- lower soil and higher L, in which transpi- tion and senescence. In P. taeda stands, ration declined rapidly between rain L in September may be twice that in April events, down to ca 15 % over 8 d without [18]. We standardized G by L so as C,day precipitation [13]. to evaluate the general seasonal pattern in without the confounding dynam- C,day G Our stand shows intermediate sensi- ics in L. This does not account for sea- tivity. Transpiration responded quickly to sonal patterns in root growth which may decreases in soil moisture availability by have a profound effect on the ability of evapotranspiration, and increased with plants to follow the water table [1, 4, 33], precipitation. Over 9 d without precipita- decouple E from current precipitation, C tion, transpiration decreased to 60 % [27], and maintain high G Thus, species . C,day a greater response than reported in Lous- with deep rooting habits can take propor- tau et al. [24] but less than in Granier and tionally more water from deep horizons Loustau [13]. Here, more frequent grow- when growing season precipitation is low ing season precipitation than in the for- [33], and there may not be a relationship mer study, and a deeper root system than between short-term soil moisture deple- in the latter study, could explain the lack of tion in the upper horizons and G but , C,day a clear seasonal pattern in G within C,day such trends may be found over the season the growing season (April-October; fig- [5]. ure 7). After October, low temperatures Under conditions in which soil mois- probably reduced G as water uptake , C,day available to plants, tran- by P. taeda was shown to be very sensitive ture reserves are spiration appears to follow potential evap- to soil temperature [22]. On a shorter time- oration [33]. In sites with deep soils scale, G of P. taeda stand in this study C,day occupied by deeply rooted species, soil showed faster response to rainless peri- moisture reserves become available ods in August and September, when L was through the process of forming new roots very high, than in May and June when L in deeper layers. In sites with shallow soils, was much lower (figure 7). Soil moisture
daytime rain) that was associated with depletion proceeds at a high rate between July to September, often than the 0.1-kPa thresh- rain events from slightly greater limiting G [44]. Using data presented old. Whether the difference in G and C,day C,day in Oren et al. [29], we calculated GC,day C,diu is positive or negative under con- G 9-d drying period and related it to ditions where rain occurs is largely depen- over a soil moisture depletion (figure 9). A 23- dent on the timing of the rain event diur- mm decrease in soil moisture of the root- nally, as daytime rain may lead to a ing zone (upper 0.35 m) nearly halved relatively larger apparent G while C,diu . C,day G nighttime rain may lead to a relatively val- larger G In general, the C,day G . C,day It is clear from viewing the difference ues (figure 7) are similar to those reported between G and G (figure 7) that C,day C,diu for other conifers [13, 14, 19, 24]. How- this difference may be positive or nega- ever, occasionally, mostly associated with tive, and is accentuated under rainy con- rain events that recharge stem storage after ditions. A positive difference indicates the a rainless period (e.g. mid-June, figure 7), influence of nighttime short- or long-term or with marginal conditions of low average recharge accounted for in G but unac- C,day D and E (e.g. late December), unreason- C counted for in G A negative differ- . C,diu may result. Just as able values of ence indicates inflated daytime values of C,day G for calculating conductance using any conductance due to low D (frequently
approach, it is necessary to evaluate the [6] Couralt D., Lagouarde J.P., Aloui B., Evap- oration for maritime catchments combining a data for unusual conditions that may cor- meteorological model with vegetation infor- rupt the calculations of In com- . C,day G mation and airborne surface temperatures, to the G calculated from the parison C,diu Agric. For. Meteorol. 82 (1996) 93-117. arithmetic mean of the diurnal pattern, Fennessey N.M., Vogel R.M., Regional mod- [7] els of potential evaporation and reference is affected less by short periods dur- C,day G evapotranspiration for the northeast USA, J. the day in which E and D are low, ing C Hydrol. 184 (1996) 337-354. when errors in measurements are high. Gash J.H.C., Stewart J.B., The average sur- [8] does not suffer from Furthermore, C,day G face resistance of a pine forest derived from Bowen-ratio measurements, Boundary Layer a systematic underestimation inherent in Meteor. 8 (1975) 453-464. G which may be substantial in days of , C,diu Gay L.W., Stewart J.B., Energy Balance stud- [9] low E and a high proportion of night- C ies in coniferous forests, Report No. 23, Instit. time water uptake. Hydrol. Natural Environ. Res. Council, Wallingford, Berkshire, 1974. Goff J.A., Gratch S., Low-pressure proper- [10] ties of water from -160 to 212 F, Trans. Am. Acknowledgements: This study was funded Soc. Heating and Ventilation Engineers 52 (1946) 95-122. by the US Department of Energy (DOE) Granier A., Une nouvelle mäthode pour la [11] through the National Institute for Global Envi- mesure de flux de säve brute dans le tronc ronmental Change (NIGEC) Southeast des arbes, Ann. Sci. For. 42 (1985) 193-200. Regional Center at the University of Alabama, Granier A., Evaluation of transpiration in a [12] Tuscaloosa (DOE Cooperative Agreement DE- Douglas fir stand by means of sap flow mea- surements, Tree Physiol. 3 (1987) 309-320. FC03-90ER61010). Granier A., Loustau D., Measuring and mod- [13] elling the transpiration of a maritime pine canopy from sap-flow data, Agric. For. Mete- orol. 71 (1994) 61-81. REFERENCES Granier A., Biron P., Bréda N., Pontailler J.- [14] Y., Saugier B., Transpiration of trees and for- est stands: short and long-term monitoring Abrams M.D., Adaptation and responses to [1] using sapflow methods, Global Change Biol. drought in Quercus species of North America, 2 (1996) 265-274. Tree Physiol. 7 (1990) 227-238. Hunt E.R., Running S.W., Federer C.A., [15] [2] Bonnefond J.M., Loustau D., Berbigier P., Extrapolating plant water flow resistances Ferreira M.I., David J.S., Pereira J.S., Tran- and capacitances to regional scales, Agric. spiration of a 64-year old maritime pine stand For. Meteorol. 54 (1991) 169-196. in Portugal: 2. Evapotranspiration and canopy Jarvis P.G., The interpretation of the varia- [16] stomatal conductance measured by an eddy tions in leaf water potential and stomatal con- covariance technique, Oecologia 107 (1996) ductance found in canopies in the field, Phil. 43-52. Trans. Roy. Soc. Lon. Ser. B 273 (1976) Bernhofer Ch., Blanford J.H., Siegwolf R. [3] 593-610. Wedler M., Applying single and two layer Jarvis P.G., McNaughton K.G., Stomatal con- [17] canopy models to derive conductances of a trol of transpiration: scaling up from leaf to Scots pine plantation from micrometeoro- region, Adv. Ecol. Res. 15 (1986) 1-49. logical measurements, Theor. Appl. Climatol. Kinerson R.S., Higginbotham K.O., Chap- [18] 53 (1996) 95-104. man R.C., The dynamics of foliage distribu- Bréda N., Granier A., Barataud F., Moyne [4] tion within a forest canopy, J. App. Ecol. 11 C., Soil water dynamics in an oak stand. I. (1974) 347-353. Soil moisture, water potentials and water Köstner B., Biron P., Siegwolf R., Granier [19] uptake by roots, Plant and Soil 172 (1995) A., Estimating water vapor flux and canopy 17-27. conductance of Scots pine at the tree level utilizing different xylem sap flow methods, Cienciala E., Lindroth A., Cermák J., Häll- [5] Theor. Appl. Climatol. 53 (1996) 105-113. gren J.-E., Kucera J., The effects of water availability on transpiration, water potential Kramer P.J., The relation between rate of [20] and growth of Picea abies during a growing transpiration and rate of absorption of water season, J. Hydrol. 155 (1994) 57-71. in plants, Am. J. Bot. 24 (1937) 10-15.
Kramer P.J., Root resistance as a cause of the [21] [35] Saugier B., Katerji N., Some plant factors absorption lag, Am. J. Bot. 25 (1938) controlling evapotranspiration, Agric. For. 110-113. Meteorol. 54 (1991) 263-268. Kramer P.J., Species differences with respect [22] Schulze E.-D., Cermak J., Matyssek R., Penka [36] absorption at low soil temperatures, to water M., Zimmermann R., Vasicek F., Gries W., Am. J. Bot. 29 (1942) 828-832. Kucera J., Canopy transpiration and water fluxes in the xylem of the trunk of Larix and [23] Kustas W.P., Stannard D.I., Allwine K.J., Picea trees - a comparison of xylem flow, Variability in surface energy flux partitioning porometer and cuvette measurements, during Washita ’92: Resulting effects on Pen- Oecologia 66 (1985) 475-483. man-Monteith and Priestly-Taylor parame- ters, Agric. For. Meteorol. 82 (1996) Shuttleworth W.J., The soil-vegetation-atmo- [37] 171-200. sphere interface, in: Raschke E., Jacob D. Loustau D., Berbigier P., Roumagnac P., [24] (Eds), Energy and Water Cycles in the Cli- Arruda-Pacheco C., David J.S., Ferreira M.I., mate System, NATO ASI Series, Vol. 5, Pereira J.S., Tavares R., Transpiration of a Springer-Verlag, Berlin, 1993, pp. 323-364. 64-year-old maritime pine stand in Portugal. Szeicz G., Long I.F., Surface resistance of [38] 1. Seasonal course of water flux through mar- crop canopies, Water Resour. Res. 5 (1969) itime pine, Oecologia 107 (1996) 33-42. 622-633. [25] McNaughton K.G., Black T.A., A study of evapotranspiration from a Douglas fir forest Taconet O., Olioso A., Ben Mehrez M., Bris- [39] using the energy balance approach, Water son N., Seasonal estimation of evaporation Resource Res. 9(6) (1973) 1579-1590. and stomatal conductance over a soybean field using surface IR temperatures, Agric. Monteith J.L., Unsworth M.H., Principles of [26] For. Meteorol. 73 (1995) 321-338. Environmental Physics, Edward Arnold, Lon- don, 1990. Tan C.S., Black T.A., Factors affecting the [40J Monteith J.L., Sceicz G., Waggoner P.E., The [27] canopy resistance of a Douglas-Fir forest, measurement and control of stomatal resis- Boundary Layer Meteor. 10 (1976) 475-488. tance in the field, J. App. Ecol. 2 (1965) Tattori S., Ikonen J.-P., Sucksdorff Y., A [41] 345-355. comparison of evapotranspiration above a Oke T.R., [28] Boundary Layer Climates, Rout- barley field based on quality tested Bowen ledge, London, New York, 1993. ratio data and Deardorff modeling, J. Hydrol. Oren R., Phillips N., Katul G., Ewers B., [29] 170 (1995) 1-14. Pataki D.E., Scaling xylem sap flux and soil [42] Sheriff D.W., Water use by Teskey R.O., water balance, and calculating variance: A Pinus radiata trees in a plantation, Tree Phys- method for partitioning water flux in forests, iol. 16 (1996) 273-279. Ann. Sci. For. (1998). Thom A.S., Oliver H.R., On Penman’s equa- [43] Pataki D.E., Oren R., Katul G., Sigmon J., [30] tion for estimating regional evapotranspira- Canopy conductance of Pinus taeda, Liq- tion, Quart. J. Roy. Meteor. Soc. 103 (1977) uidambar styraciflua and Quercus phellos 345-357. under varying atmospheric and soil moisture conditions, Tree Physiol. (1998). Todd P.H., Use of a local mass balance [44] [31] Phillips N., Oren R., Zimmermann R., Radial method to determine the effects of soilwater patterns of xylem sap flow in non-, diffuse- conditions on canopy conductance and to esti- and ring-porous tree species, Plant Cell Env- mate forest stand evapotranspiration in a nat- iron. 19 (1996) 983-990. ural forest and in growth chamber, MS The- sis, Duke University, Durham, NC, 1995. [32] Phillips N., Nagchaudhuri A., Oren R., Katul G.G., Time constant for water uptake in Wallace J.S., Calculating evaporation: resis- [45] loblolly pine trees estimated from time series factors, Agric. For. Meteorol. 73 tance to of stem sapflow and evaporative demand, (1995) 353-366. Trees (1998). Whitehead D., Teskey R.O., Dynamic [46] Rambal S., Water balance and pattern of root [33] response of stomata to changing irradiance water uptake by a Quercus coccifera L. ever- in loblolly pine (Pinus taeda L.), Tree Phys- green scrub, Oecologia 62 (1984) 18-25. iol. 15 (1995) 245-251. Rochette P., Desjardins R.L., Dwyer L.M., [34] Stewart D.W., Pattey E., Dube P.A., Estima- Zhang L., Dawes W.R., Hatton T.J., Mod- [47] tion of maize (Zea mays L.) canopy conduc- elling hydrologic processes using a biophys- tance by scaling up leaf stomatal conduc- ically based model - application of WAVES tance, Agric. For. Meteorol. 54 (1991) to FIFE and HAPEX-MOBILHY, J. Hydrol. 241-262. 185 (1996) 147-169.