Patterns of plant rehydration and growth following pulses ...

Biogeosciences 18 831ndash847 2021httpsdoiorg105194bg-18-831-2021copy Author(s) 2021 This work is distributed underthe Creative Commons Attribution 40 License

Patterns of plant rehydration and growth following pulses of soilmoisture availabilityAndrew F Feldman1 Daniel J Short Gianotti1 Alexandra G Konings2 Pierre Gentine3 and Dara Entekhabi11Department of Civil and Environmental Engineering Massachusetts Institute of Technology 15 Vassar StCambridge Massachusetts 02139 USA2Department of Earth System Science Stanford University Stanford California USA3Department of Earth and Environmental Engineering Columbia University New York New York USA

Correspondence Andrew F Feldman (afeld24mitedu)

Received 13 October 2020 ndash Discussion started 16 October 2020Revised 9 December 2020 ndash Accepted 14 December 2020 ndash Published 5 February 2021

Abstract Plant hydraulic and photosynthetic responses toindividual rain pulses are not well understood because fieldexperiments of pulse behavior are sparse Understanding in-dividual pulse responses would inform how rainfall intermit-tency impacts terrestrial biogeochemical cycles especially indrylands which play a large role in interannual global atmo-spheric carbon uptake variability Using satellite-based esti-mates of predawn plant and soil water content from the SoilMoisture Active Passive (SMAP) satellite we quantify thetimescales of plant water content increases following rain-fall pulses which we expect bear the signature of whole-plant mechanisms In wetter regions we find that plant wa-ter content increases rapidly and dries along with soil mois-ture which we attribute to predawn soilndashplant water poten-tial equilibrium Global drylands by contrast show multi-day plant water content increases after rain pulses Shorterincreases are more common following dry initial soil condi-tions These are attributed to slow plant rehydration due tohigh plant resistances using a plant hydraulic model Longermulti-day dryland plant water content increases are attributedto pulse-driven growth following larger rain pulses and wet-ter initial soil conditions These dryland responses reflectwidespread drought recovery rehydration responses and in-dividual pulse-driven growth responses as supported by pre-vious isolated field experiments The response dependenceon moisture pulse characteristics especially in drylands alsoshows ecosystem sensitivity to intra-annual rainfall intensityand frequency which are shifting with climate change

1 Introduction

A changing climate is likely to shift not only mean annualrainfall but also the frequency and intensity of rainfall events(Donat et al 2016 Giorgi et al 2019 Trenberth 2011) Un-derstanding the impacts of rainfall distribution shifts on theterrestrial biosphere is important because vegetation globallyexerts a large control on terrestrial water and carbon balances(Ahlstroumlm et al 2015 Jasechko et al 2013 Poulter et al2014) and drives feedbacks with the lower atmosphere (Gen-tine et al 2019 Green et al 2017) Changing rainfall fre-quency with the same annual rainfall can impact terrestrialcarbon uptake (Fay et al 2003 Knapp et al 2002) sug-gesting ecosystem sensitivity to characteristics of individualrain events This motivates characterizing plant responses toindividual moisture availability pulses across climates andbiomes This is especially the case for semi-arid herbaceousplants which respond primarily to individual pulses likelyoccurring under a pulse reserve paradigm of individual rain-fall events triggering photosynthetic responses and storages(Collins et al 2014 Feldman et al 2018 Noy-Meir 1973)However vegetation responses on these shorter timescalesare less well understood than the more commonly studied re-sponses to monthly or annual water anomalies

A major component of understanding moisture pulse re-sponses is quantifying the duration over which plants takeup and use the rain pulse water These response durationsbear the signature of among others plant rehydration in thesoilndashplantndashatmosphere continuum (SPAC) growth and re-source limitations such as drought (Manzoni et al 2013

Published by Copernicus Publications on behalf of the European Geosciences Union

832 A F Feldman et al Patterns of plant rehydration

Ogle et al 2015 Ogle and Reynolds 2004 Sperry et al2016) Characterizing these timescales and their dependen-cies across biomes will increase our understanding of whole-plant behavior and assist plant hydraulic parameterizations inland surface models to better assess plant water stress (Bonanet al 2014 Fisher et al 2018 Kennedy et al 2019 Lin etal 2019 Tai et al 2017 Xu et al 2016) However the fun-damentals of these pulse water use durations especially soil-to-plant water storage timescales remain unknown globally

Soil and plant water content measurements can be used tocharacterize the soilndashplant hydraulic system and understandplant water storage timescales but they are laborious and arethus often constrained to a single location Alternatively mi-crowave remote sensing satellites provide plant water contentobservations across the globe at near-daily sampling frequen-cies (Entekhabi et al 2010 Kerr et al 2010 Konings et al2016) Although these observations are at a coarse resolution(tens of kilometers) their ecosystem-scale resolution and re-lationship to leaf water potential (Momen et al 2017 Zhanget al 2019) make them a useful tool for studies of ecosystemplantndashwater relations (Feldman et al 2018 Konings et al2019 Konings and Gentine 2017) However the plant watercontent measurements are a function of both relative watercontent and dry biomass thus making them additionally sen-sitive to biomass and growth (Momen et al 2017 Zhanget al 2019) This is nevertheless an advantage because theplant water content sensitivity to both water potential and drybiomass allows evaluation of timescales of multiple whole-plant mechanisms at the landscape scale

With regard to plant rehydration timescales plant wateruptake and storage timescales have typically been assessedusing an electric circuit analogy with the timescale of in-terest being the plant resistance times capacitance or RCtime constant (Phillips et al 1997 2004 Ward et al 2013)The RC time constant quantifies the time required for leafor xylem water potential to reach 63 of its equilibriumvalue following a soil moisture or transpiration perturbationMeasured RC time constants vary from minutes for grassesto hours for trees (Hunt and Nobel 1987 Nobel and Jor-dan 1983 Phillips et al 1997 2004 Ward et al 2013)According to this theory plant rehydration after rain pulsesshould occur within a day By contrast field pulse experi-ments of grass and shrub species primarily in dryland envi-ronments (broadly annual rainfall less than 500 mm) com-monly show multi-day predawn water potential increasesafter rewetting pulses (Fravolini et al 2005 Huxman etal 2004 Ignace et al 2007 West et al 2007) Multi-daypredawn water potential increases would lengthen plant wa-ter content timescales These multi-day water potential in-creases appear to be related to recovery from water limita-tion between storms which highlights the potential impactof antecedent moisture conditions on plant responses (Guoand Ogle 2019 Ogle et al 2015 Plaut et al 2013) Hy-draulic limitations from previously dry conditions have alsobeen observed driving multi-day recovery of leaf gas ex-

change after soil rewetting (Blackman et al 2009 Brodribband Cochard 2009 Chen et al 2009 Huxman et al 2004Martorell et al 2014) These multi-day hydraulic responseobservations call into question whether hydraulic responsetimescales are consistently sub-daily across global biomesThey also highlight the unknown role of moisture pulse char-acteristics (antecedent soil conditions and pulse magnitude)on these timescales

With regard to timescales of growth while growth isknown to occur on seasonal timescales there is evidencethat growth can occur following rainfall events as under thepulse reserve hypothesis (Noy-Meir 1973) Specifically dry-land measurements suggest that growth can occur over daysto weeks following a pulse (Angert et al 2007 Doughertyet al 1996 Hermance et al 2015 Novoplansky and Gold-berg 2001 Post and Knapp 2019 Sher et al 2004) Alsoecosystem growth responses in drylands have been mod-eled previously with a 1ndash5 d lag (depending on plant type)and a decaying persistence over 1ndash2-week scales (Ogle andReynolds 2004 Reynolds et al 2004) Ultimately pulse-driven growth following rainfall would lengthen plant watercontent timescales by increasing the total plant water storagecapacity

Here we evaluate the duration of total plant water con-tent increases following rainfall pulse events Under nom-inal moisture conditions with no growth one would ex-pect sub-daily plant water content increases based on RCtime constants Slow rehydration andor growth would likelyextend these timescales to multiple days We ask acrossglobal biomes do plant water uptake responses to soilmoisture pulses ever occur beyond a day and what arethese timescales How do pulse characteristics (pulse mag-nitude moisture pre-conditions) and growth influence thesetimescales Do attributes of the moisture pulse (pulse mag-nitude moisture pre-conditions) favor plant growth versusrehydration To address these questions we use microwaveremote sensing of total plant water content a combinationof dry biomass and relative water content following an ap-proach for rain pulse studies originally developed in Feld-man et al (2018) In order to better understand potentialmechanisms underlying remote-sensing-observed timescalevariations we also discuss the observed timescales and theirdrivers in the context of a SPAC model

2 Methods

21 Datasets

We use 4 years (1 April 2015 to 31 March 2019) of soil mois-ture and plant water content observations from the Soil Mois-ture Active Passive (SMAP) satellite (Entekhabi et al 2010)SMAP measures the low-frequency microwave (14 GHz) ra-diation emitted from Earthrsquos surface The radiation signal isin units of temperature or brightness temperature (TB) The

Biogeosciences 18 831ndash847 2021 httpsdoiorg105194bg-18-831-2021

A F Feldman et al Patterns of plant rehydration 833

radiation is polarized where the emitted wavesrsquo oscillationshave distinct horizontal (TBH) and vertical (TBV) orienta-tion SMAP measures both TBV and TBH Both TBV andTBH magnitudes alone are sensitive to surface soil moisture(top sim 5 cm) Furthermore the difference between TBV andTBH is sensitive to how much the emitted waves are atten-uated when traversing a vegetation canopy The vegetationattenuation of the microwave radiation is called vegetationoptical depth (VOD) More vegetation water content resultsin higher VOD (Jackson and Schmugge 1991 Konings etal 2019) An established radiative transfer equation can par-tition the TBV and TBH signals into soil moisture and VOD(Mo et al 1982 Wigneron et al 2017) We use a recentlydeveloped algorithm called the multi-temporal dual channelalgorithm (MT-DCA) to robustly estimate soil moisture andVOD using this radiative transfer equation (Konings et al2016 2017)

The SMAP satellite measurements occur at 0600 (localtime) everywhere on 9 km grids across the globe The time0600 is approximately predawn when plant water status isassumed to be maximal (due to nighttime plant rehydration)The satellite orbit is such that there is a 1 2 or 3 d revisitdepending on the day and latitude Due to the orbit patternhigher latitudes are measured more frequently This resultsin sampling frequencies of 1ndash2 d at midlatitudes and 2ndash3 d atthe Equator

Since VOD has been shown to be nearly linearly pro-portional to total vegetation water content (Jackson andSchmugge 1991) VOD is proportional to the product ofrelative water content and aboveground dry biomass (Kon-ings et al 2019 Momen et al 2017 Zhang et al 2019)Therefore VOD can increase due to either rehydration of cellwater storages or biomass growth as growth provides addi-tional water storage capacity VOD is expected to be sensi-tive to rehydration because of near-linear relationships be-tween relative water content and plant water potential es-pecially for herbaceous species which are primarily inves-tigated in this study (Jones 2014 Jones and Higgs 1979Konings et al 2019 Nobel and Jordan 1983) While thelow resolution of VOD estimates hinders species-specific orstand-scale assessments it provides the opportunity to assessintegrated landscape-scale vegetation behavior across globalbiomes (Feldman et al 2018 Tian et al 2018) VOD showspromise for use in monitoring plant water stress with recentfindings showing VOD can monitor time evolution of plantwater stress and drought-induced mortality with loss of plantwater storage (Feldman et al 2020 Martiacutenez-Vilalta et al2019 Rao et al 2019)

Soil moisture observations from the MT-DCA algorithmcompare closely to other SMAP soil moisture products(which use different algorithms) as well as to in situ obser-vations (Chan et al 2016 Dadap et al 2019 Feldman etal 2018) Direct in situ VOD information is unavailable al-though SMAP VODrsquos mean and dynamics are comparable toanother satellite VOD product (Kerr et al 2010) For further

discussion of SMAP VOD estimate performance and com-parison with other products we refer the reader to Koningset al (2017) and Feldman et al (2018)

To assist in discriminating VOD changes related to hy-draulic or growth activity we use the daily leaf area index(LAI) product from the Spinning Enhanced Visible and In-frared Imager (SEVIRI) on board EUMETSATrsquos MeteosatSecond Generation (MSG-2) satellite series (Trigo et al2011a) These LAI observations serve as an indicator forabove-ground biomass independent of VOD While con-strained primarily to Africa these LAI observations are es-timated from 15 min geostationary observations which pro-vide daily LAI fluctuations after cloud contamination mit-igation Both VOD and LAI datasets together are requiredto determine the occurrence of pulse-driven growth VODincreases can be linked directly to a specific rain event be-cause of SMAPrsquos more rapid effective sampling (due to nocover contamination) but they are confounded by rehydra-tion LAI changes over the weekly scales of pulses can detectcanopy growth but because of a non-linear averaging tech-nique (Garciacutea-Haro and Camacho 2014) the LAI datasetis partially smoothed over sub-weekly scales and may beless apt to determine whether detected growth over a weekis specifically associated with a given rain event Neverthe-less increasing LAI over a rain event can identify whetherVOD increases associated with that storm are due to growthor only rehydration As such we are interested in using LAIchanges qualitatively to determine whether LAI is increasingor decreasing over more than week-long periods Thereforebiases in LAI magnitude do not influence the analysis

Use of SEVIRI LAI for this application is preferred dueto SEVIRIrsquos frequent sampling and filtering techniques thatprovide better resolution of the seasonal growth and senes-cence stages especially during the wet season than otheravailable satellites (Garciacutea-Haro et al 2013 Gessner etal 2013) SEVIRI ultimately provides 3ndash5 d effective sam-pling during cloud-contaminated periods which is typically4 times less than effective sampling with low-Earth-orbitsatellites (ie MODIS) (Fensholt et al 2006) Thereforedespite being global low-Earth-orbit satellites are not usedbecause they sample too coarsely in time for the applica-tions here Furthermore SEVIRI LAI retrievals in the herba-ceous biomes evaluated in Africa have the lowest retrievalerrors (Garciacutea-Haro et al 2013) Therefore SEVIRI LAI islikely to detect increasing biomass over 1- to 2-week peri-ods While no other adequate satellites exist for direct com-parison with results here the analysis was repeated with theSEVIRI fraction of absorbed photosynthetically active radia-tion (FAPAR) observations derived from different measure-ment frequencies than LAI and similar results were obtained(Fig S1)

Ancillary data are used to evaluate climate and biome de-pendencies of the findings Specifically we compute meanannual precipitation using the Global Precipitation Mea-surement IMERG product (Huffman 2015) and tree cover

httpsdoiorg105194bg-18-831-2021 Biogeosciences 18 831ndash847 2021

Figure 1 Schematic definitions of rain pulse soil moisture dry-down period and time to peak plant water content (tp)

from the Moderate Resolution Imaging Spectroradiometer(MODIS) (Dimiceli et al 2015) to evaluate VOD behavioracross climate gradients International Geosphere-BiosphereProgramme (IGBP) land cover maps are used to removefrozen and bare ground (Kim 2013) Tree cover is also usedto remove densely vegetated forests where soil moisture andVOD satellite estimation are uncertain

22 Soil moisture pulse identification

VOD behavior and response timescales are evaluated dur-ing soil moisture drydown periods that occur after rainfall(Fig 1) Drydown periods or soil moisture pulses are de-fined as an increase in soil moisture of at least 001 m3m3

followed by a drying period of at least four consecutive mea-surements (approximately 8ndash12 d) This approach is nearlyidentical to previous approaches (McColl et al 2017 Shel-lito et al 2018) To remove seasonal drydowns less asso-ciated with an individual rainfall event drying periods oflonger than 20 d are not included Seasonal trends (periodicclimatology) are removed from soil moisture and VOD timeseries during drydowns while preserving the magnitude ofinitial conditions (Feldman et al 2019) This procedure re-moves seasonal VOD growth trends to isolate short-term in-creases associated with a given storm that are due to pulse-driven growth andor slow rehydration Note that differencespersist in the literature on whether the rain event or plant re-sponse is defined as the pulse (Reynolds et al 2004) Bothsoil and plant responses are discussed as pulses here

23 Vegetation pulse response timescale estimation andanalysis

For infiltrating water to ultimately reach the leaf it must typi-cally percolate from the soil surface to the roots pass throughthe root endodermis and move up the xylem through theshoot to leaves (Mackay et al 2015 Sperry et al 19982016) Given an identified soil moisture drydown periodthe VOD response timescale defined here as time to peak(tp) is estimated as the time from the beginning of the soil

moisture drydown to the first local maximum value of VOD(Fig 1) After a period of water storage plant water con-tent loss occurs due to surface and atmospheric drying andwarming creating a peak in plant water content conditionsthat tp attempts to quantify (Feldman et al 2020) tp cap-tures the aggregate rehydration and growth timescale duringthis soilndashplant water transport process The tp estimation re-lies on consecutive VOD increases which provide more ro-bust estimates than the global maximum during the drydownTo increase sample size we conduct the analysis on all pixelscontained within a 05times05 domain (includessim 30 SMAPpixels) Densely forested regions (gt 40 tree cover suchas the Congo and Amazon basins) are masked because soilmoisture and VOD estimates are less certain from radiativetransfer limitations in dense canopies (Feldman et al 2018Konings et al 2017)

We compute the median tp over all drydowns within each05times 05 pixel The tp probability mass function withina given pixel typically has a mixed distribution with manyzeros resembling a zero-inflated Poisson distribution Themedian tp is chosen to describe this distribution because itnot only provides a typical timescale of VOD increase butalso indicates whether or not the majority of pulses resultedin consecutive multi-day VOD increases (as opposed to themean which can be greater than zero even if a majority ofpulses resulted in a tp of zero) Several tests are performed todetermine the effects of SMAPrsquos irregular above-daily sam-pling period the algorithm and measurement noise on tp es-timates for a given pulse (see Sect 24)

The tp definition evaluates continuous post-rainfall VODincreases and potentially neglects the duration of plant watercontent increases during the period of soil moisture increase(between the observations before the drydown beginning andat the drydown beginning) We do not attempt to estimate theduration of VOD increase during the soil moisture increaseperiod because it is not possible to resolve when plant watercontent increases initiated due to the 1ndash3 d satellite samplingfrequency Instead the VOD behavior preceding the drydownis categorically evaluated by determining the frequency ofplant water content increases during the rain pulse This al-lows evaluations of tp of zero which can result from eithera rapid rehydration response during the rainstorm (on the or-der of hours) or no rehydration response throughout the pulse(no VOD increase)

For each soil moisture pulse within a pixel tp is esti-mated along with the LAI change from beginning to endof the drydown (1LAI) antecedent surface soil moisture(soil moisture value before drydown beginning) soil mois-ture pulse magnitude (difference between initially pulsed andantecedent surface soil moisture) and antecedent VOD An-tecedent is defined here as the observation just precedingthe peak soil moisture observation beginning the drydownEach variable is binned into rapid VOD response (tp = 0)short VOD increase (1le tp le 3 d) and long VOD increase(tp gt 3 d) groups because they provide partitions consistent

with the satellite sampling and because uncertainty analysesreveal that while a tp estimate for a given drydown is un-certain there is more confidence in whether it exists withina given bin (see Sect 34) The groups of three differenttp lengths are then compared for each respective metric of1LAI antecedent surface soil moisture soil moisture pulsemagnitude and antecedent VOD Due to non-normality ofgroups based on JarquendashBera normality tests (Jarque andBera 1980) KruskalndashWallis non-parametric tests are per-formed to determine significance of difference in mediansbetween the tp groups for each respective metric Also cor-relation coefficients are computed between tp and1LAI an-tecedent moisture and pulse magnitude to augment the cate-gorical analyses

The seasonal timing of rapid short and long tp values isassessed relative to peak seasonal moisture or the proxim-ity to the wet season The peak seasonal soil moisture is de-termined by smoothing the soil moisture times series usinga 90 d moving-average window This only provides a zero-order seasonal moisture peak approximation as many loca-tions have intermittent rainfall or bimodal precipitation dis-tributions

24 Satellite plant water content response uncertaintyanalysis

Several tests were conducted to evaluate the robustness of tpestimates given uncertainties due to a 1ndash3 d satellite samplingfrequency the soil moisturendashVOD retrieval algorithm andrandom instrument noise on the order of that of the SMAPradiometer (Piepmeier et al 2017) A stochastic rainfall gen-erator was used to simulate soil moisture and consequentdrydowns A range of ldquotruerdquo VOD behavior was consideredsuch as perfect correlation with soil moisture (true tp of zero)and multi-day VOD increases during drying (true tp greaterthan zero) Analyses were conducted directly on these simu-lated time series including converting these time series to TBmeasurements for implementation in the algorithm and com-paring the original true VOD time series to the algorithm-estimated VOD time series as in Zwieback et al (2019) Fortests with the 1ndash3 d satellite sampling frequency the effect ofrandomly removing observations every 1ndash2 d on tp was as-sessed To test the effect of the soil moisturendashVOD retrievalalgorithm on tp tp was estimated after inputting true TB mea-surements into the retrieval algorithm Finally to assess theeffect of instrument noise on tp estimates this aforemen-tioned process was repeated by adding normally distributedrandom error to TB measurements

25 Plant hydraulic model simulations

To investigate the underlying mechanisms that alter plantrehydration timescales we evaluate plant hydraulic storagetimescales under varying conditions after a surface soil mois-ture pulse using a plant hydraulic model We specificallychoose a one-dimensional soilndashplantndashatmosphere continuum(SPAC) model assessed in previous studies (Carlson andLynn 1991 Hartzell et al 2017 Lhomme et al 2001Zhuang et al 2014) Note that assimilating satellite VODinto a SPAC model is beyond the scope of this study and ishindered by the large number of unknown plant hydraulic pa-rameters at global scales SPAC simulations are repeated andrandomized using a Monte Carlo approach drawing from pa-rameter distributions based on previous field measurementsMore details can be found about the SPAC model in the SI

3 Results

31 Global plant water content characteristic responsesand timescales

The VOD data show that more arid regions with lower an-nual rainfall and tree cover (Fig 2b and c) exhibit multi-dayvegetation water content increases following moisture pulses(tp ge 1 d blue regions in Fig 2a) That is after soil mois-ture increases following a storm vegetation water contentincreases for multiple days even while surface soil moisturebegins to dry Furthermore in regions with tp ge 1 d VODtypically begins increasing during the rain pulse period in-stead of with a lag after soil moisture drying begins (occursin 77 of the pixels) Aggregated example time series ofthis nonzero tp behavior can be seen in drylands in the Saheland southwest United States (Fig 3a and b) In the regionswith multi-day VOD increases the spatial median tp is 2 dNote that various responses are spatially aggregated togetherto produce the post-rainfall responses in Fig 3 In subsequentsections we evaluate and partition the mechanisms underly-ing these multi-day plant water content increases primarilyin drylands (blue regions in Fig 2a)

By contrast more humid ecosystems with more woodyplant coverage typically do not exhibit multi-day plant watercontent increases (tp = 0 Fig 2) They instead exhibit waterloss following the pulse during soil drying (see average be-havior illustrated in Fig 3c and d) In 83 of regions with tpof zero (red regions in Fig 2a) the plant drying responses aretypically preceded by an initial VOD increase showing rapidwater uptake during the storm period (Fig S2) In contrast aminority of these regions typically show no VOD increasessuggesting plant water content continuously dries throughoutthe pulse with no discernable hydraulic response (Fig S2)We do not investigate regions with median tp of zero furtherhere because their exact sub-daily timescales are unresolv-able but within expectations (see Discussion)

Figure 2 Median time to peak plant water content (tp) after soil moisture pulse (a) Median tp global distribution Median tp binned as afunction of (b) mean annual precipitation and (c) tree cover Mostly bare surfaces with low vegetation density are masked Densely forestedareas (tree cover gt 40 ) are masked due to limitations in VOD estimation for dense canopies

Figure 3 VOD rate of change distribution on a given day after the pulse for regions outlined in the insets Boxes delineate the interquartilerange for each day dVOD dt is normalized by dividing by VOD time mean for a given pixel for consistent comparison across regionsdVOD dt is reported as the average change rate over a given day (for example from day 0 to day 1) All pixels with the noted dominantland cover (gt 75 IGBP land cover type) are used within the boxed region in the inset to create the distributions for each respective dayafter the pulse Gray shading indicates the pulse period when soil moisture is increasing (Fig 1) At time greater than zero soil moisture isdrying (drydown event see Fig 1) Behavior extends beyond a week in many cases but only 8 d following the pulse are shown here Notethat top and bottom panels have different y-axis limits

Figure 4 Relationship of plant water content increase timescaleswith biomass changes in African regions with median tp ge 1 dGrowth increases the timescale of plant water content MannndashWhitney U tests indicate that the medians of the two bins are sig-nificantly different (p lt 005)

32 Growth influence on plant water content increasetimescales

A positive correlation between LAI rates of change andplant water content increase timescales is found in 72 of African pixels with median tp ge 1 (p lt 005) There-fore longer tp values are associated with increasing biomasswithin a given pixel (Fig 4) Calculating the LAI rates ofchange for the rapid VOD response (tp = 0) short VOD in-crease (1le tp le 3) and long VOD increase (tp gt 3) groupsreveals that growth tends to occur alongside plant water con-tent increases longer than 3 d (Fig 5a c and d) These longerplant water content uptake timescales average 7 d and con-tinue beyond a week 40 of the time This growth influencemeans that rehydration alone cannot explain longer plantwater content increase durations Note that VOD increasesduring growth still demonstrate increased aboveground plantwater content because more aboveground biomass requireswater uptake to hydrate a greater volumetric plant storagecapacity There are some pixels that show declining biomassduring longer tp (Fig 5d) We attribute these cases to detec-tion of longer tp during senescence in regions where senes-cence of leaf area is differentially more rapid than growthUltimately we interpret overall spatial patterns and avoidinterpreting individual pixels acknowledging noisy tp esti-mates in some cases (see Sect 34)

In general growth does not influence shorter plant watercontent increase timescales LAI is often decreasing whentp is 1ndash3 d (Fig 5a) Therefore plant water content increasesover less than 3 d are mostly due to rehydration Furthermorewhen VOD increases do not extend beyond a day (tp = 0)growth is also less frequently occurring

The reoccurrence of growth-influenced multi-day VODincreases consistently following soil moisture pulses meansthat rainfall intermittently triggers growth throughout a yeartp values greater than 3 d are linked to pulse-driven growthbecause they coincide with increasing daily LAI (Fig 5)

consistently co-occur with a soil moisture pulse and are sep-arated from seasonal growth patterns Our seasonal detrend-ing of VOD isolates these pulsed plant growth responsesfrom seasonal growth cycles These isolated sub-weeklyVOD responses closely link to the timing of moisture pulsessuggesting a causendasheffect of rain pulse followed by plant wa-ter content response

Although this daily LAI dataset is limited to Africa onlyAfrica contains one-third of the worldrsquos regions with mediantp ge 1 d (blue regions in Fig 2a) and we expect similar re-sults for the rest of the globe Note that these results are notsensitive to the 3 d threshold choice between long and shortVOD increase groups they are nearly identical if choosing athreshold of 2 4 or 5 d Furthermore results repeated withFAPAR are qualitatively the same (Fig S1 see Sect 21)

On average the short and long VOD increase bins occurapproximately with equal frequency both with seasonal vari-ations (Fig 5b) Longer-duration VOD increases influencedby growth (Fig 5a) appear to occur more frequently duringtimes of the year when soil moisture is higher (Fig 5b) Incontrast short VOD increases associated more with rehydra-tion occur more often during drier times of the year (Fig 5b)Furthermore rapid rehydration responses occur 40 ndash50 of the time throughout the year amongst the multi-day VODincreases

LAI growth rates average 0005 m2m2 per day for theselong VOD increases On a mean percent change basis thistranslates to a 15 LAI increase on average over the courseof a week after a pulse Note that LAI may not detect ad-ditional branchndashstem biomass growth that VOD may detectUltimately we are more interested in qualitatively increas-ing trends in LAI rather than the magnitudes of LAI rates ofchange which are less certain

33 Pulse condition influence on plant water contentincrease timescales

Variations in VOD increase timescales across space and timelikely occur as a result of differences in vegetation traitsedaphic and topographic properties affecting soil moistureinfiltration and climatic properties While an evaluation ofall of these factors is beyond the scope of this paper we fo-cus here on climatic drivers To evaluate the climatic driversof VOD increase timescales in regions with median tp ge

1 d (blue regions in Fig 2a) we assess how tp relates torain pulse conditions antecedent surface soil moisture soilmoisture pulse magnitude and antecedent VOD Growth-influenced VOD increases of longer duration are associatedwith initially wetter surface soil (Fig 6a) as well as withlarger pulse magnitudes (Fig 6b) This suggests that the sur-face must be sufficiently wet initially and a large enoughpulse must occur to elicit a growth response Converselyshorter-duration VOD increases associated primarily withrehydration frequently occur under drier initial soil condi-tions with smaller rewetting pulses (Fig 6) This is consistent

Figure 5 Timescale of plant water content increases in relation to biomass changes and seasonality in African regions with median tp ge 1 dGrowth influences the plant water uptake timescale when 1LAI 1t gt 0 By contrast only rehydration contributes to plant water contentincreases when1LAI 1t lt 0 Only intermittent variability in VOD is used to produce tp removing confounding seasonal connections withLAI (see text and SI) (a) Mean change in LAI per day over length of pulse period binned into rapid responses (tp = 0) short VOD increases(1le tp le 3 d) and long VOD increases (tp gt 3 d) A KruskalndashWallis test indicates group medians are all significantly different (p 001χ2= 2576 υ = 2) Pairwise MannndashWhitney U tests confirm that all pairs are significantly different (p lt 005) (b) Seasonality of short and

long VOD increase occurrences with respect to seasonal soil moisture peak Positive and negative time indicates occurrence after and beforethe soil moisture seasonal peak respectively Plotted values are spatial medians in 60 d sized bins Sample size in each bin (in a given pixel)is over 100 though pulses tend to be more frequent closer to seasonal soil moisture peak (c) Spatial distribution of median 1LAI 1t forshort VOD increases as binned in (a) (d) Spatial distribution of median 1LAI 1t for long VOD increases as binned in (a)

with short increase durations becoming more prevalent dur-ing drier periods and long increase durations becoming moreprevalent in wet periods (Fig 5b) Note that while these re-sults are shown globally they are nearly identical when cal-culated for only Africa (not shown) and therefore they canbe consistently compared with the growth assessment resultsand timescale bins (Sect 32 Fig 5)

In assessing what differentiates rapid responses (tp = 0 d)and short VOD increases (tp = 1ndash3 d) that appear driven byonly rehydration we find short VOD increases have slightlylarger pulse magnitudes (Fig 6b) and drier antecedent soilmoisture than rapid responses (Fig 6a) Also drier initialplant water status for short VOD increases (Fig 6c) inde-pendently suggests a slightly drier root zone initially than forrapid responses (Fig S13) Note that mean differences aresmall between these metrics even though they show statisti-cal significance (likely effect of large sample size deflating pvalues) Nevertheless cases of vegetation water content in-crease on the order of 1ndash3 d due primarily to rehydrationoccur under dry soil conditions with small to moderate rewet-ting pulses

Satellite tp estimates appear robust with effects of satellitesampling frequency algorithmic estimation error and mea-surement noise increasing tp variance but not introducingdiscernable biases The SMAP sampling period of 1ndash3 d re-sults in greater variance but no mean biases for tp estimatesbelow the Nyquist frequency of 4ndash6 d (Figs S4 and S5)One can combine low-frequency microwave measurementsfrom similar satellites (Kerr et al 2010) to increase the sam-pling frequency and reduce uncertainty in tp estimates hereThis is not attempted due to complications in combining thedatasets The MT-DCA algorithm used here reduces sensitiv-ity to noise within the simultaneous soil moisturendashVOD es-timation (Konings et al 2015 2016 Zwieback et al 2019)We found that use of a traditional algorithm biases tp towardszero (Fig S7) because its greater sensitivity to noise will tendto spuriously induce positive correlation between soil mois-ture and VOD within the estimation procedure (Konings etal 2016) Therefore increases in VOD during soil drying

Figure 6 Global spatial distribution of pulse conditions binned as a function of rapid VOD response (tp = 0) short VOD increases (tp = 1ndash3 d) and long VOD increases (tp gt 3 d) in regions with median tp ge 1 d KruskalndashWallis (KW) tests indicate all group medians are sig-nificantly different within each panel and pairwise MannndashWhitney U tests confirm that all possible combinations of differences in groupmedians across (a) (b) and (c) are significantly different (p lt 005) (a) Antecedent surface soil moisture (KW test p 001 χ2

= 2200υ = 2) A total of 77 of pixels have significantly positive linear relationships with tp (p lt 005) (b) Surface soil moisture pulse magni-tude (KW test p 001 χ2

= 7819 υ = 2) A total of 85 of pixels have significantly positive linear relationships with tp (p lt 005)(c) Antecedent VOD (KW test p 001 χ2

=163 υ = 2) A total of 81 of pixels have significantly negative linear relationships with tp(p lt 005)

and thus positive tp values are not a result of algorithmic ar-tifacts from the MT-DCA algorithm used here (Feldman etal 2018) It is also unlikely that algorithmic noise is drivingspatial patterns as both algorithms produce the same tp spa-tial patterns Note that the MT-DCA algorithm can slightlyartificially increase tp though measurement noise may can-cel this effect (Fig S4) Finally measurement noise primarilyincreases the variance of tp (Fig S4)

Ultimately while identifying precise tp values for a givendrydown may be hindered by these sources of uncertaintymedian tp values for a pixel are likely not biased and moreconfidence is exhibited in whether tp is zero or non-zero(Fig S6) This uncertainty analysis provides confidence inthe global patterns of median tp and results based on binnedtp where zero short and long tp can be confidently parti-tioned

4 Discussion

41 Plant water uptake timescale variation acrossclimates

We observe a continuum of plant water uptake timescalesfrom humid to dryland environments with mainly drylandsshowing frequent multi-day plant water content increases af-ter rainfall before water loss occurs (Fig 2) Given that planthydraulic capacitance increases at least 3 orders of magni-tude from grasses in drylands to trees in humid regions (Carl-son and Lynn 1991 Hunt et al 1991) one might expect ifat all occurrence of multi-day responses in wooded regionsHowever humid wooded regions broadly exhibit peak plantwater content during rather than after the storm event beforesoil drying begins (Figs 2 and S2) Plant water loss occursthereafter (Fig 3c and d) likely due to simultaneous soil andplant drying where plant rehydration becomes progressivelyrestricted with drying soil (Feldman et al 2020) The ini-tial VOD increase can be due to plant water uptake wherepre-dawn water potential approaches equilibrium with soilmoisture andor due to plant interception of rainfall droplets

In some cases no discernible VOD increase occurs before orafter the pulse which may indicate sufficiently well-wateredconditions (Fig S2) Even in drylands pulse water utiliza-tion for plant rehydration decreases if the plantndashsoil systemis initially sufficiently wet (Ehleringer et al 1991 Gebaueret al 2002 Ignace et al 2007) Nevertheless due to the 1ndash3 d satellite sampling we are unable to resolve more specificplant water content timescales and underlying mechanismsfor these well-watered wooded regions

The consistent trend of multi-day plant water content in-creases which are found broadly across dry regions (Fig 2)is unexpected at least in the context of nominal RC timeconstants (plant water uptake and storage timescales) Field-based estimates of plant water uptake timescales (via RCplant hydraulic time constants) typically do not exceed a dayregardless of species (Huang et al 2017 Nobel and Jordan1983 Phillips et al 1997 2004 Ward et al 2013) This is inpart because plant capacitance and resistance tend to trade offwith changes in plant architecture and moisture conditions(ie capacitance increases and resistance decreases gener-ally from grass to tree species) (Hunt et al 1991 Phillipset al 1997 Richards et al 2014 Ward et al 2013) Wefind both the influence of growth and slow plant rehydrationcontribute to these observed multi-day VOD increases Wediscuss these growth and plant rehydration mechanisms ob-served in drylands further below

42 Growth impact on dryland plant water uptaketimescales

As is evident in independent satellite LAI observationsgrowth increases the duration of plant water content in-creases (Fig 4) and appears to occur primarily for plant wa-ter content increases of more than 3 d in dryland regions(Fig 5) These week-long consecutive plant water contentincreases occur when the soil is initially wetter and pulsesare larger (Fig 6) These results are based on 1ndash2-week in-creasing trends in LAI coinciding with VOD increases ofmore than 3 d Confidence is exhibited in these sub-monthlyLAI trends because of SEVIRIrsquos ability to resolve the sea-sonal growth stages during the wet season lower LAI un-certainty in Africarsquos biomes with herbaceous vegetation andSEVIRIrsquos filtering of LAI noise Therefore plant rehydrationalone cannot explain these longer-duration VOD increasesWe further suspect rehydration is rapid under these well-watered conditions While pulsed growth is expected to oc-cur with a lag of 1ndash5 d (Ogle and Reynolds 2004) theselags may be obscured in the sampling of VOD and initialVOD increases due to rehydration Furthermore these pulsedplant water content increases due to growth may continuefor longer than detected here (beyond 2 weeks) Howevercontinued water loss and VOD decreases through transpi-ration may eventually dominate over VOD increases dueto growth curtailing the peak VOD (resulting in behaviorlike that shown schematically in Fig 1) VOD ultimately

shows sub-weekly growth temporal dynamics beyond thoseresolved from optical instruments

These results indicate that large soil moisture pulses oninitially wetter soils trigger dryland vegetation growth re-sponses after storm events as hypothesized under the pulsereserve paradigm (Collins et al 2014 Noy-Meir 1973)This weekly variability at least in part drives seasonalgrowth in these locations (Reynolds et al 1999) wherethe seasonal growth cycles appear to be made up of sub-weekly intermittent growth dynamics as modeled in Ogleand Reynolds (2004) The growth occurrences under wetterconditions are expected given that cell turgor must be highfor cell expansion and rapid growth to occur (Kramer andBoyer 1995) Furthermore a recent study showed that largerpulses during the growing season resulted in 1ndash2 weeks ofincreasing leaf and stalk density in a semi-arid grasslandconsistent with results here (Post and Knapp 2019) Addi-tionally larger pulses have previously been shown to elicitgreater plant photosynthetic responses (Chen et al 2009Dougherty et al 1996 Schwinning and Sala 2004) In asimilar study these longer satellite-based plant water uptakeresponses coincided with larger and longer carbon uptake re-sponses at dryland flux tower sites following larger moisturepulses on initially wet soils (Feldman et al 2021) There-fore detection of pulse-triggered growth on timescales ofdrydowns here is consistent with previous results althoughit is the first to show how widespread the pulse-triggeredgrowth dynamics are in drylands Additionally the seasonaloccurrence of growth-driven longer tp (Fig 5b) supports thefact that pulses will trigger growth primarily in the seasonwhen species are phenologically active and able to invest inaboveground biomass (Post and Knapp 2019 Reynolds etal 1999 Schwinning and Sala 2004)

43 Slow dryland plant rehydration mechanisms

Over half of the moisture pulses primarily in global dry-lands result in multi-day satellite-observed plant water con-tent increases (Fig 2) These multi-day VOD increases areoften only due to rehydration especially the shorter VODincreases (1ndash3 d) following small to moderate pulses on ini-tially dry soils (Figs 5 and 6) They can occur even whenbiomass is decreasing (Fig 5c such as leaf off) where therelative water content increases are larger than what the VODincrease signal suggests For dryland ecosystems that in-clude grass and shrub species with isolated forests multi-day rehydration is generally unexpected with nominal RCtime constants on the order of an hour (Carlson and Lynn1991 Hunt et al 1991) However previous field studies of-ten show 1ndash4 d rehydration of grasses and shrubs upon rewet-ting following dry conditions especially in the southwesternUnited States where multi-day VOD increases are observed(Briones et al 1998 Fravolini et al 2005 Huxman et al2004 Ignace et al 2007 West et al 2007)

Figure 7 SPAC model simulations of determined sufficient conditions driving slow rehydration (see text and SI) for semi-arid grass andshrub species Rate of change in predawn water potential (ψw) of all plant water stores on a given day following a pulse where dψw dt gt 0indicates rehydration Same format and conventions as Fig 3 Parameter bounds determined to drive each slow rehydration scenario areshown in each panel (a) Plant limitation only where plant resistance (Rp) is initially high and decreases (b) Soil limitation only where rootzone soil moisture is initially dry and pulses are small to moderate causing slow infiltration (c) Both plant and soil limitations superposedfrom (a) and (b) Parameter ranges common amongst all simulations rooting depth= 03 to 07 m vapor pressure deficit (VPD)= 1 to 5 kPawind= 1 to 8 ms capacitance= 10minus6 to 10minus5 mMPa RS impairment factor=minus10 tominus1 See SI for more information on the SPAC modeland simulations

To better understand the physiological drivers of multi-day rewetting we assessed the potential hydrologic andphysiological mechanisms driving slow rehydration using aplant hydraulic (SPAC) model and parameters within knownbounds for semi-arid species (Figs S8 to S14 and Table S1)We find that the sufficient conditions for multi-day plant re-hydration determined here include initially high soilndashplantresistances decreasing over multiple days following a stormThese time-varying resistances can occur either in the soilplant or both (Figs 7 S9 and S11) The possibility ofmulti-day rehydration due to these conditions suggests thatRC timescales can greatly deviate from nominal conditions(Scholz et al 2011) especially under drought scenarioswhere resistances are both higher and changing

After uncoupling effects of soil and plant resistances in theSPAC model we suspect that multi-day rehydration as seenby VOD is dominated by plant resistance limitations ratherthan soil resistance limitations This is because high soil re-sistances reduce infiltration rates and result in a phase-laggeddelay in plant rehydration (Fig 7b) which is not observed

in the satellite VOD behavior here In the slow rehydrationcases (tp = 1ndash3 d) VOD increases begin immediately dur-ing the storm and not with a phase-lagged delay (Fig S3)This behavior more closely resembles slow plant rehydrationdominated by plant resistance limitations rather than thosedominated by soil resistance limitations For example 1ndash3 duptake timescales based on satellite VOD observations ap-pear like that in Fig 3a and b which more closely resembleSPAC model simulations in Fig 7a than in Fig 7b Note thatboth conditions may be present within a coarse-resolutionpixel because the pixel spatially averages plant water con-tent behavior over the landscape As a result a combina-tion of behaviors like those in Fig 7 aggregate into the spa-tially averaged behavior like that shown in Fig 3a and bTherefore while plant resistance limitations may dominatemost landscapes that show 1ndash3 d VOD increases based onthe above discussion slow infiltration responses may still bespatially prevalent with a potential dependence on sub-pixelantecedent moisture variability

The initially high decreasing resistances as determinedfrom the SPAC model and likely influencing landscape-scaleplant water content behavior are likely due to drought recov-ery of the soilndashroot interface and xylem architecture Initiallyhigh decreasing plant resistances have been observed in thefield where after rewetting of dry soil conditions soilndashrootinterface and xylem resistances can decrease by 1 to 3 or-ders of magnitude over a few days (Carminati et al 2017North and Nobel 1995 Trifilograve et al 2004 West et al 2007)Under prolonged dry conditions a disconnect between soiland root interface can occur and after rewetting the soilndashroot and radial root hydraulic conductivity progressively in-crease (Carminati et al 2009 North and Nobel 1997) Sim-ilarly xylem cavitation and embolism from drying lead toincreased xylem resistance that can regain conductance andrefill after rewetting (Martorell et al 2014) though notingcontroversies with existence of xylem repair and refilling(Charrier et al 2016 Lamarque et al 2018 Venturas etal 2017) Recent evidence suggests that whole-root resis-tance (ie soilndashroot interface radial) rather than xylem re-sistance (from cavitation) dominates the whole-plant resis-tance during these drying and rewetting cycles (Rodriguez-Dominguez and Brodribb 2020) Finally fine root growthcan occur after rewetting which can contribute to decreasingroot resistances though these effects may occur over longerweekly scales (Eissenstat et al 1999)

5 Conclusions

The globally observed timescales of plant water content re-sponses to moisture pulses here reveal a climate gradientof ecosystem-scale vegetation pulse water usage The veg-etation water content of more wooded humid regions ap-pears to respond rapidly to rain pulses likely with rehydra-tion responses occurring in less than a day (due to predawnequilibrium) By contrast drier ecosystems more often showmulti-day plant water uptake responses following moisturepulses with the timescale of the response indicative of un-derlying mechanisms Specifically longer plant water con-tent increases are linked to growth and follow larger pulseson wetter surfaces Therefore dryland vegetation intermit-tently upregulates and grows after individual rainfall eventsdemonstrating spatially extensive evidence for the pulse re-serve hypothesis Specifically we show that there is a com-ponent of growth linked directly to individual rainfall eventsin addition to any continuous seasonal growth (Noy-Meir1973) Additionally shorter plant water content increasesare indicative of slow plant rehydration responses and arelinked here to hydraulic recovery from initially dry condi-tions The slow rehydration responses indicate that plant wa-ter uptake timescales can frequently deviate from nominalRC time constants with greatly increased resistances underdry conditions as observed previously in field experimentsand demonstrated here using a SPAC model

Our results also indicate that SMAP satellite vegetationoptical depth observations hold biophysical information atsub-weekly timescales Namely they show patterns of rehy-dration growth responses and rain pulse dependencies con-sistent with that seen in field studies These satellite-basedplant water content responses were also shown to have simi-lar response signatures to carbon uptake responses at drylandfield sites (Feldman et al 2021) This merits investigation ofsub-monthly ecological processes using these 1ndash3 d sampledsatellite microwave observations which so far have been pri-marily used for seasonal and interannual VOD variability in-vestigations (Brandt et al 2018 Jones et al 2014 Tian etal 2018)

We demonstrate that global dryland ecosystems exhibita high sensitivity to the characteristics of individual mois-ture pulses Therefore expected shifts in rainfall frequencyand intensity may influence arid to semi-arid vegetation hy-draulic and growth processes presenting potential feedbackson biogeochemical cycles and changes in plant communitycomposition (Giorgi et al 2019 Knapp et al 2002) Thesedry ecosystems cover 40 of the land surface store signif-icant amounts of carbon (Beer et al 2010 Collins et al2014) regulate atmospheric carbon interannual variability(Ahlstroumlm et al 2015 Poulter et al 2014) and are projectedto expand (Huang et al 2016) Therefore it is key to charac-terize the vegetation responses to rainfall events ndash includingtheir timescales ndash in these environments in the context of pre-dicting future climate

Code availability The SPAC model used in the studywas created by the first author and can be accessedat httpsgithubcomafeld24VOD_TimescalesblobmasterFeldmanetal2021_SPACTimeSeriesm (last access15 January 2021 Feldman 2020) All scripts and re-lated data files displaying the figures are available athttpsgithubcomafeld24VOD_Timescales (last access 15 Jan-uary 2021 Feldman 2020)

Data availability SMAP L1C brightness temperatures used to re-trieve soil moisture are available from the National Snow and IceData Center (NSIDC) (httpsnsidcorgdataSPL1CTB_E last ac-cess 5 September 2020 Chaubell et al 2016) LandSAF leafarea index is available from EUMETSAT (httpslandsafipmaptenproductsvegetationlai last access 15 November 2020 Trigoet al 2011b) Generated maps are available at httpsgithubcomafeld24VOD_Timescales (last access 15 January 2021 Feldman2020)

Supplement The supplement related to this article is available on-line at httpsdoiorg105194bg-18-831-2021-supplement

Author contributions PG and AFF conceived the study DE ledthe project AFF conducted the analysis and wrote the manuscriptDJSG AGK PG and DE contributed interpretations and numerousrevisions to all versions of the manuscript analysis and figures

Competing interests The authors declare that they have no conflictof interest

Special issue statement This article is part of the special is-sue ldquoMicrowave remote sensing for improved understanding ofvegetation-water interactions (BGHESS inter-journal SI)rdquo It is aresult of the EGU General Assembly 2020 3ndash8 May 2020

Acknowledgements The authors thank Missy HolbrookTony Rockwell Anju Manandhar and Jess Gersony of theHolbrook Plant Physiology Laboratory at Harvard Universityfor many insightful discussions The authors also thank the twoanonymous reviewers for their insightful comments

Financial support This research has been supported by theNational Aeronautics and Space Administration (grant nos1510842 80NSSC18K0715 NNH19ZDA001N-SMAP) and theNational Oceanic and Atmospheric Administration (grant noNA17OAR4310127)

Review statement This paper was edited by Martin De Kauwe andreviewed by two anonymous referees

References

Ahlstroumlm A Raupach M R Schurgers G Smith B ArnethA Jung M Reichstein M Canadell J G FriedlingsteinP Jain A K Kato E Poulter B Sitch S Stocker B DViovy N Wang Y P Wiltshire A Zaehle S and ZengN The dominant role of semi-arid ecosystems in the trendand variability of the land CO2 sink Science 348 895ndash900httpsdoiorg1010022015JA021022 2015

Angert A L Huxman T E Barron-Gafford G A Gerst K Land Venable D L Linking growth strategies to long-term pop-ulation dynamics in a guild of desert annuals J Ecol 95 321ndash331 httpsdoiorg101111j1365-2745200601203x 2007

Beer C Reichstein M Tomelleri E Ciais P Jung M Carval-hais N Roumldenbeck C Arain M A Baldocchi D Bonan GB Bondeau A Cescatti A Lasslop G Lindroth A LomasM Luyssaert S Margolis H Oleson K W Roupsard OVeenendaal E Viovy N Williams C Woodward F I andPapale D Terrestrial gross carbon dioxide uptake Global dis-tribution and covariation with climate Science 329 834ndash838httpsdoiorg101126science1184984 2010

Blackman C J Brodribb T J and Jordan G J Leaf hydraulicsand drought stress Response recovery and survivorship in four

woody temperate plant species Plant Cell Environ 32 1584ndash1595 httpsdoiorg101111j1365-3040200902023x 2009

Bonan G B Williams M Fisher R A and Oleson K WModeling stomatal conductance in the earth system linking leafwater-use efficiency and water transport along the soil-plant-atmosphere continuum Geosci Model Dev 7 2193ndash2222httpsdoiorg105194gmd-7-2193-2014 2014

Brandt M Wigneron J P Chave J Tagesson T PenuelasJ Ciais P Rasmussen K Tian F Mbow C Al-Yaari ARodriguez-Fernandez N Schurgers G Zhang W Chang JKerr Y Verger A Tucker C Mialon A Rasmussen LV Fan L and Fensholt R Satellite passive microwaves re-veal recent climate-induced carbon losses in African drylandsNat Ecol Evol 2 827ndash835 httpsdoiorg101038s41559-018-0530-6 2018

Briones O Montantildea C and Ezcurra E International Associa-tion for Ecology Competition Intensity as a Function of ResourceAvailability in a Semiarid Ecosystem Oecologia 116 365ndash3721998

Brodribb T J and Cochard H Hydraulic failure defines the recov-ery and point of death in water-stressed conifers Plant Physiol149 575ndash584 httpsdoiorg101104pp108129783 2009

Carlson T N and Lynn B The effects of plant water stor-age on transpiration and radiometric surface temperature AgricFor Meteorol 57 171ndash186 httpsdoiorg1010160168-1923(91)90085-5 1991

Carminati A Vetterlein D Weller U Vogel H J and OswaldS E When roots lose contact Vadose Zone J 8 898ndash809httpsdoiorg102136vzj20080147 2009

Carminati A Benard P Ahmed M A and Zarebanadkouki MLiquid bridges at the root-soil interface Plant Soil 417 1ndash15httpsdoiorg101007s11104-017-3227-8 2017

Chaubell J Chan S Dunbar R S Peng J and Yueh S SMAPL1C enhanced brightness temperatures available at httpsnsidcorgdataSPL1CTB_E (last access 5 September 2020)2016

Chan S K Bindlish R OrsquoNeill P E Njoku E Jackson TColliander A Chen F Burgin M Dunbar S Piepmeier JYueh S Entekhabi D Cosh M H Caldwell T Walker JWu X Berg A Rowlandson T Pacheco A McNairn HThibeault M Martinez-Fernandez J Gonzalez-Zamora ASeyfried M Bosch D Starks P Goodrich D Prueger JPalecki M Small E E Zreda M Calvet J C Crow WT and Kerr Y Assessment of the SMAP Passive Soil Mois-ture Product IEEE Trans Geosci Remote Sens 54 4994ndash5007httpsdoiorg101109TGRS20162561938 2016

Charrier G Torres-Ruiz J M Badel E Burlett R Choat BCochard H Delmas C E L Domec J C Jansen S KingA Lenoir N Martin-StPaul N Gambetta G A and DelzonS Evidence for hydraulic vulnerability segmentation and lackof xylem refilling under tension Plant Physiol 172 1657ndash1668httpsdoiorg101104pp1601079 2016

Chen S Lin G Huang J and Jenerette D Depen-dence of carbon sequestration on the differential responsesof ecosystem photosynthesis and respiration to rain pulsesin a semiarid steppe Glob Change Biol 15 2450ndash2461httpsdoiorg101111j1365-2486200901879x 2009

Collins S L Belnap J Grimm N B Rudgers J A DahmC N DrsquoOdorico P Litvak M Natvig D O Peters D

C Pockman W T Sinsabaugh R L and Wolf B O AMultiscale Hierarchical Model of Pulse Dynamics in Arid-Land Ecosystems Annu Rev Ecol Evol Syst 45 397ndash419httpsdoiorg101146annurev-ecolsys-120213-091650 2014

Dadap N C Cobb A R Hoyt A M Harvey C F and KoningsA G Satellite soil moisture observations predict burned areain Southeast Asian peatlands Environ Res Lett 14 094014httpsdoiorg1010881748-9326ab3891 2019

Dimiceli C Carroll M Sohlberg R Kim D H KellyM and Townshend J R G MOD44B MODISTerra Veg-etation Continuous Fields Yearly L3 Global 250m SINGrid V006 2015 NASA EOSDIS Land Processes DAAChttpsdoiorg105067MODISMOD44B006 2015

Donat M G Lowry A L Alexander L V OrsquoGormanP A and Maher N More extreme precipitation in theworldrsquos dry and wet regions Nat Clim Change 6 508ndash513httpsdoiorg101038nclimate2941 2016

Dougherty R L Lauenroth W K and Singh J S Response ofa Grassland Cactus to Frequency and Size of Rainfall Eventsin a North American Shortgrass Steppe J Ecol 84 177httpsdoiorg1023072261353 1996

Ehleringer J R Phillips S L Schuster W S F and SandquistD R Differential utilization of summer rains by desert plantsOecologia 88 430ndash434 httpsdoiorg101007BF003175891991

Eissenstat D M Whaley E L Volder A and WellsC E Recovery of citrus surface roots following pro-longed exposure to dry soil J Exp Bot 50 1845ndash1854httpsdoiorg101093jxb503411845 1999

Entekhabi D Njoku E G OrsquoNeill P E Kellogg K HCrow W T Edelstein W N Entin J K Goodman SD Jackson T J Johnson J Kimball J Piepmeier J RKoster R D Martin N McDonald K C Moghaddam MMoran S Reichle R Shi J C Spencer M W Thur-man S W Tsang L and Van Zyl J The Soil MoistureActive Passive (SMAP) Mission Proc IEEE 98 704ndash716httpsdoiorg101109JPROC20102043918 2010

Fay P A Carlisle J D Knapp A K Blair J M andCollins S L Productivity responses to altered rainfall pat-terns in a C 4-dominated grassland Oecologia 137 245ndash251httpsdoiorg101007s00442-003-1331-3 2003

Feldman A F Generated vegetation optical depth timescaledatasets available at httpsgithubcomafeld24VOD_Timescales (last access 15 January 2021) 2020

Feldman A F Short Gianotti D J Konings A G McColl K AAkbar R Salvucci G D and Entekhabi D Moisture pulse-reserve in the soil-plant continuum observed across biomes NatPlants 4 1026ndash1033 httpsdoiorg101038s41477-018-0304-9 2018

Feldman A F Short Gianotti D J Trigo I F Salvucci G Dand Entekhabi D Satellite-Based Assessment of Land SurfaceEnergy Partitioning-Soil Moisture Relationships and Effects ofConfounding Variables Water Resour Res 55 10657ndash10677httpsdoiorg1010292019WR025874 2019

Feldman A F Short Gianotti D J Trigo I F Salvucci GD and Entekhabi D Land-atmosphere drivers of landscape-scale plant water content loss Geophys Res Lett 47e2020GL090331 httpsdoiorg1010292020GL090331 2020

Feldman A F Chulakadabba A Short Gianotti D J andEntekhabi D Landscape-scale plant water content and car-bon flux behavior following moisture pulses From drylandto mesic environments Water Res 57 e2020WR027592httpsdoiorg1010292020WR027592 2021

Fensholt R Sandholt I Stisen S and Tucker C AnalysingNDVI for the African continent using the geostationary meteosatsecond generation SEVIRI sensor Remote Sens Environ 101212ndash229 httpsdoiorg101016jrse200511013 2006

Fisher R A Koven C D Anderegg W R L Christoffersen BO Dietze M C Farrior C E Holm J A Hurtt G C KnoxR G Lawrence P J Lichstein J W Longo M Matheny AM Medvigy D Muller-Landau H C Powell T L Serbin SP Sato H Shuman J K Smith B Trugman A T ViskariT Verbeeck H Weng E Xu C Xu X Zhang T and Moor-croft P R Vegetation demographics in Earth System Models Areview of progress and priorities Glob Change Biol 24 35ndash54httpsdoiorg101111gcb13910 2018

Fravolini A Hultine K R Brugnoli E Gazal R English NB and Williams D G Precipitation pulse use by an invasivewoody legume The role of soil texture and pulse size Oecolo-gia 144 618ndash627 httpsdoiorg101007s00442-005-0078-42005

Garciacutea-Haro F J and Camacho F Algorithm Theoretical Ba-sis Document for Vegetation parameters (VEGA) Ref NumberSAFLANDUVVR_VEGA20 Issue 20 2014

Garciacutea-Haro F J Camacho F and Meliaacute J The EUMETSATSatellite Application Facility on Land Surface Analysis Prod-uct User Manual Vegetation Parameters (VEGA) Ref NumberSAFLANDUVVR_VEGA_MSG Issue 31 2013

Gebauer R L E Schwinning S and Ehleringer J R Inter-specific Competition and Resource Utilization between Bumble-bees Ecology 83 2602ndash2616 httpsdoiorg10230736720072002

Gentine P Green J K Gueacuterin M Humphrey V Seneviratne SI Zhang Y and Zhou S Coupling between the terrestrial car-bon and water cycles ndash a review Environ Res Lett 14 083003httpsdoiorg1010881748-9326ab22d6 2019

Gessner U Niklaus M Kuenzer C and Dech S Intercompar-ison of leaf area index products for a gradient of sub-humid toarid environments in west africa Remote Sens 5 1235ndash1257httpsdoiorg103390rs5031235 2013

Giorgi F Raffaele F and Coppola E The response of precipita-tion characteristics to global warming from climate projectionsEarth Syst Dynam 10 73ndash89 httpsdoiorg105194esd-10-73-2019 2019

Green J K Konings A G Alemohammad S H BerryJ Entekhabi D Kolassa J Lee J E and GentineP Regionally strong feedbacks between the atmosphereand terrestrial biosphere Nat Geosci 10 410ndash414httpsdoiorg101038ngeo2957 2017

Guo J S and Ogle K Antecedent soil water contentand vapor pressure deficit interactively control water po-tential in Larrea tridentata New Phytol 221 218ndash232httpsdoiorg101111nph15374 2019

Hartzell S Bartlett M S and Porporato A The roleof plant water storage and hydraulic strategies in rela-tion to soil moisture availability Plant Soil 419 503ndash521httpsdoiorg101007s11104-017-3341-7 2017

Hermance J F Augustine D J and Derner J D Quan-tifying characteristic growth dynamics in a semi-arid grass-land ecosystem by predicting short-term NDVI phenol-ogy from daily rainfall a simple four parameter coupled-reservoir model Int J Remote Sens 36 5637ndash5663httpsdoiorg1010800143116120151103916 2015

Huang C W Domec J C Ward E J Duman T Manoli GParolari A J and Katul G G The effect of plant water storageon water fluxes within the coupled soil-plant system New Phy-tol 213 1093ndash1106 httpsdoiorg101111nph14273 2017

Huang J Yu H Guan X Wang G and Guo R Accelerateddryland expansion under climate change Nat Clim Change 6166ndash171 httpsdoiorg101038nclimate2837 2016

Huffman G GPM Level 3 IMERG Final Run Half Hourly 01times01Degree Precipitation version 05 NASA Goddard Space FlightCenter Active Archive Center GSFC DAAC 2015

Hunt E R and Nobel P S Non-steady-state Water Flow for ThreeDesert Perennials with Different Capacitances Aust J PlantPhysiol 14 363ndash375 1987

Hunt Jr E R Running S W and Federer C A Extrapolatingplant water flow resistances and capacitances to regional scalesAgric For Meteorol 54 169ndash195 1991

Huxman T E Cable J M Ignace D D Eilts J A En-glish N B Weltzin J and Williams D G Responseof net ecosystem gas exchange to a simulated precipitationpulse in a semi-arid grassland The role of native versusnon-native grasses and soil texture Oecologia 141 295ndash305httpsdoiorg101007s00442-003-1389-y 2004

Ignace D D Huxman T E Weltzin J F and Williams DG Leaf gas exchange and water status responses of a na-tive and non-native grass to precipitation across contrastingsoil surfaces in the Sonoran Desert Oecologia 152 401ndash413httpsdoiorg101007s00442-007-0670-x 2007

Jackson T J and Schmugge T J Vegetation effects on the mi-crowave emission of soils Remote Sens Environ 36 203ndash212httpsdoiorg1010160034-4257(91)90057-D 1991

Jarque C M and Bera A K Efficient test for normality ho-moscedasticity and serial independence of regression residualsEcon Lett 6 255ndash259 1980

Jasechko S Sharp Z D Gibson J J Birks S J Yi Y andFawcett P J Terrestrial water fluxes dominated by transpira-tion Nature 496 347ndash350 httpsdoiorg101038nature119832013

Jones H G Plants and Microclimate A Quantitative Approach toEnvironmental Plant Physiology 3rd ed Cambridge UniversityPress Cambridge UK 2014

Jones H G and Higgs K H Water potential-water con-tent relationships in apple leaves J Exp Bot 30 965ndash970httpsdoiorg101093jxb305965 1979

Jones M O Kimball J S and Nemani R R AsynchronousAmazon forest canopy phenology indicates adaptation to bothwater and light availability Environ Res Lett 9 124021httpsdoiorg1010881748-9326912124021 2014

Kennedy D Swenson S Oleson K W Fisher R ALawrence D M da Costa A C L and Gentine PImplementing plant hydraulics in the Community LandModel version 5 J Adv Model Earth Syst 1ndash29httpsdoiorg1010292018ms001500 2019

Kerr Y Waldteufel P Wigneron J-P Delwart S Cabot FBoutin J Escorihuela M J Font J Reul N Gruhier CJuglea S E Drinkwater M R Achim Hreul N Boutin JGruhier C Juglea S E Hahne A Neira M M and Meck-lenburg S The SMOS Mission New Tool for Monitoring KeyElements of the Global Water Cycle Proc IEEE 98 666ndash6872010

Kim S Ancillary Data Report Landcover Classification Califor-nia Institute of Technology SMAP Science Document no 042D-53057 2013

Knapp A K Fay P A Blair J M Collins S L Smith M DCarlisle J D Harper C W Danner B T Lett M S andMcCarron J K Rainfall variability carbon cycling and plantspecies diversity in a mesic grassland Science 298 2202ndash2205httpsdoiorg101126science1076347 2002

Konings A G and Gentine P Global variations in ecosystem-scale isohydricity Glob Change Biol 23 891ndash905httpsdoiorg101111gcb13389 2017

Konings A G McColl K A Piles M and Entekhabi D Howmany parameters can be maximally estimated from a set of mea-surements IEEE Geosci Remote Sens Lett 12 1081ndash1085httpsdoiorg101109LGRS20142381641 2015

Konings A G Piles M Rotzer K McColl K A Chan SK and Entekhabi D Vegetation optical depth and scatteringalbedo retrieval using time series of dual-polarized L-band ra-diometer observations Remote Sens Environ 172 178ndash189httpsdoiorg101016jrse201511009 2016

Konings A G Piles M Das N and Entekhabi D L-bandvegetation optical depth and effective scattering albedo esti-mation from SMAP Remote Sens Environ 198 460ndash470httpsdoiorg101016jrse201706037 2017

Konings A G Rao K and Steele-Dunne S C Macroto micro microwave remote sensing of plant water contentfor physiology and ecology New Phytol 223 1166ndash1172httpsdoiorg101111nph15808 2019

Kramer P J and Boyer J S Water Relations of Plants and SoilsAcademic Press San Diego CA USA 1995

Lamarque L J Corso D Torres-Ruiz J M Badel E BrodribbT J Burlett R Charrier G Choat B Cochard H Gam-betta G A Jansen S King A Lenoir N Martin-StPaulN Steppe K Van den Bulcke J Zhang Y and Delzon SAn inconvenient truth about xylem resistance to embolism in themodel species for refilling Laurus nobilis L Ann For Sci 7588 httpsdoiorg101007s13595-018-0768-9 2018

Lhomme J P Rocheteau A Ourcival J M and Rambal SNon-steady-state modelling of water transfer in a Mediterraneanevergreen canopy Agric For Meteorol 108 67ndash83 2001

Lin C Gentine P Frankenberg C Zhou S Kennedy D andLi X Evaluation and mechanism exploration of the diurnal hys-teresis of ecosystem fluxes Agric For Meteorol 278 107642httpsdoiorg101016jagrformet2019107642 2019

Mackay D S Roberts D E Ewers B E Sperry J S McDow-ell N G and Pockman W T Interdependence of chronic hy-draulic dysfunction and canopy processes can improve integratedmodels of tree response to drought Water Resour Res 516156ndash6176 httpsdoiorg1010022015WR017200A 2015

Manzoni S Vico G Porporato A and Katul G Bio-logical constraints on water transport in the soil-plant-

atmosphere system Adv Water Resour 51 292ndash304httpsdoiorg101016jadvwatres201203016 2013

Martiacutenez-Vilalta J Anderegg W R L Sapes G and Sala AGreater focus on water pools may improve our ability to under-stand and anticipate drought-induced mortality in plants NewPhytol 223 22ndash32 httpsdoiorg101111nph15644 2019

Martorell S Diaz-Espejo A Medrano H Ball M Cand Choat B Rapid hydraulic recovery in Eucalyptuspauciflora after drought Linkages between stem hydraulicsand leaf gas exchange Plant Cell Environ 37 617ndash626httpsdoiorg101111pce12182 2014

McColl K A Wang W Peng B Akbar R Short Gianotti DJ Lu H Pan M and Entekhabi D Global characterization ofsurface soil moisture drydowns Geophys Res Lett 44 3682ndash3690 httpsdoiorg1010022017GL072819 2017

Mo T Choudhury B J Schmugge T J Wang J Rand Jackson T J A model for microwave emissionfrom vegetation-covered fields J Geophys Res 87 11229httpsdoiorg101029JC087iC13p11229 1982

Momen M Wood J D Novick K A Pangle R Pock-man W T McDowell N G and Konings A G Interact-ing Effects of Leaf Water Potential and Biomass on Vegeta-tion Optical Depth J Geophys Res-Biogeo 122 3031ndash3046httpsdoiorg1010022017JG004145 2017

Nobel P S and Jordan P W Transpiration stream ofdesert species Resistances and capacitances for a c3a c4 and a cam plant J Exp Bot 34 1379ndash1391httpsdoiorg101093jxb34101379 1983

North G B and Nobel P S Hydraulic conductivity of concentricroot tissues of Agave deserti Engelm under wet and drying con-ditions New Phytol 130 47ndash57 httpsdoiorg101111j1469-81371995tb01813x 1995

North G B and Nobel P S Root-soil contact for the desert suc-culent Agave deserti in wet and drying soil New Phytol 13521ndash29 httpsdoiorg101046j1469-8137199700620x 1997

Novoplansky A and Goldberg D E Effects of water pulsing onindividual performance and competitive hierarchies in plants JVeg Sci 12 199ndash208 httpsdoiorg1023073236604 2001

Noy-Meir I Desert Ecosystems Environment and ProducersAnnu Rev Ecol Syst 4 25ndash52 1973

Ogle K and Reynolds J F Plant responses to precip-itation in desert ecosystems Integrating functional typespulses thresholds and delays Oecologia 141 282ndash294httpsdoiorg101007s00442-004-1507-5 2004

Ogle K Barber J J Barron-Gafford G A Bentley L P YoungJ M Huxman T E Loik M E and Tissue D T Quantifyingecological memory in plant and ecosystem processes Ecol Lett18 221ndash235 httpsdoiorg101111ele12399 2015

Phillips N Nagchaudhuri A Oren R and Katul G Time con-stant for water transport in loblolly pine trees estiamted fromtime series of evaporative demand and stem sapflow Trees 11412ndash419 1997

Phillips N G Oren R Licata J and Linder S Time series di-agnosis of tree hydraulic characteristics Tree Physiol 24 879ndash890 httpsdoiorg101093treephys248879 2004

Piepmeier J R Focardi P Horgan K A Knuble J EhsanN Lucey J Brambora C Brown P R Hoffman P JFrench R T Mikhaylov R L Kwack E Y Slimko EM Dawson D E Hudson D Peng J Mohammed P N

De Amici G Freedman A P Medeiros J Sacks F Es-tep R Spencer M W Chen C W Wheeler K B Edel-stein W N OrsquoNeill P E and Njoku E G SMAP L-Band Microwave Radiometer Instrument Design and First Yearon Orbit IEEE Trans Geosci Remote Sens 55 1954ndash1966httpsdoiorg101109TGRS20162631978 2017

Plaut J A Wadsworth W D Pangle R Yepez E A McdowellN G and Pockman W T Reduced transpiration response toprecipitation pulses precedes mortality in a pintildeon-juniper wood-land subject to prolonged drought New Phytol 200 375ndash387httpsdoiorg101111nph12392 2013

Post A K and Knapp A K Plant growth and above-ground production respond differently to late-season del-uges in a semi-arid grassland Oecologia 191 673ndash683httpsdoiorg101007s00442-019-04515-9 2019

Poulter B Frank D Ciais P Myneni R B Andela N Bi JBroquet G Canadell J G Chevallier F Liu Y Y RunningS W Sitch S and Van der Werf G R Contribution of semi-arid ecosystems to interannual variability of the global carbon cy-cle Nature 509 600ndash603 httpsdoiorg101038nature133762014

Rao K Anderegg W R L Sala A Martiacutenez-Vilalta J andKonings A G Satellite-based vegetation optical depth as an in-dicator of drought-driven tree mortality Remote Sens Environ227 125ndash136 httpsdoiorg101016jrse201903026 2019

Reynolds J F Virginia R A Kemp P R De Soyza A Gand Tremmel D C Impact of drought on desert shrubs Ef-fects of seasonality and degree of resource island develop-ment Ecol Monogr 69 69ndash106 httpsdoiorg1018900012-9615(1999)069[0069IODODS]20CO2 1999

Reynolds J F Kemp P R Ogle K and Fernaacutendez R J Mod-ifying the ldquopulse-reserverdquo paradigm for deserts of North Amer-ica Precipitation pulses soil water and plant responses Oecolo-gia 141 194ndash210 httpsdoiorg101007s00442-004-1524-42004

Richards A E Wright I J Lenz T I and Zanne A E Sapwoodcapacitance is greater in evergreen sclerophyll species growingin high compared to low-rainfall environments Funct Ecol 28734ndash744 httpsdoiorg1011111365-243512193 2014

Rodriguez-Dominguez C M and Brodribb T J Declin-ing root water transport drives stomatal closure in oliveunder moderate water stress New Phytol 225 126ndash134httpsdoiorg101111nph16177 2020

Scholz F G Phillips N G Bucci S J Meinzer F C and Gold-stein G Size- and Age-Related Changes in Tree Structure andFunction in Size- and Age-Related Changes in Tree Structureand Function vol 4 2011

Schwinning S and Sala O E Hierarchy of responses to resourcepulses in arid and semi-arid ecosystems Oecologia 141 211ndash220 httpsdoiorg101007s00442-004-1520-8 2004

Shellito P J Small E E and Livneh B Controls on surfacesoil drying rates observed by SMAP and simulated by the Noahland surface model Hydrol Earth Syst Sci 22 1649ndash1663httpsdoiorg105194hess-22-1649-2018 2018

Sher A A Goldberg D E and Novoplansky A The effect ofmean and variance in resource supply on survival of annuals fromMediterranean and desert environments Oecologia 141 353ndash362 httpsdoiorg101007s00442-003-1435-9 2004

Sperry J S Adler F R Campbell G S and Comstock J PLimitation of plant water use by rhizosphere and xylem conduc-tance Results from a model Plant Cell Environ 21 347ndash359httpsdoiorg101046j1365-3040199800287x 1998

Sperry J S Wang Y Wolfe B T Mackay D S An-deregg W R L McDowell N G and Pockman WT Pragmatic hydraulic theory predicts stomatal responsesto climatic water deficits New Phytol 212 577ndash589httpsdoiorg101111nph14059 2016

Tai X Mackay D S Anderegg W R L Sperry J S andBrooks P D Plant hydraulics improves and topography me-diates prediction of aspen mortality in southwestern USA NewPhytol 213 113ndash127 httpsdoiorg101111nph14098 2017

Tian F Wigneron J-P Ciais P Chave J Ogeacutee J PentildeuelasJ Raeligbild A Domec J-C Tong X Brandt M Mialon ARodriguez-Fernandez N Tagesson T Al-Yaari A Kerr YChen C Myneni R B Zhang W Ardouml J and Fensholt RCoupling of ecosystem-scale plant water storage and leaf phe-nology observed by satellite Nat Ecol Evol 2 1428ndash1435httpsdoiorg101038s41559-018-0630-3 2018

Trenberth K E Changes in precipitation with climate changeClim Res 47 123ndash138 httpsdoiorg103354cr00953 2011

Trifilograve P Raimondo F Nardini A Lo Gullo M A andSalleo S Drought resistance of Ailanthus altissima Roothydraulics and water relations Tree Physiol 24 107ndash114httpsdoiorg101093treephys241107 2004

Trigo I F Dacamara C C Viterbo P Roujean J Ole-sen F Barroso C Camacho-de-coca F Freitas S CGarciacutea-haro J Geiger B Ghilain N Meliaacute J PessanhaL and Arboleda A The Satellite Application Facility forLand Surface Analysis Int J Remote Sens 1161 2725ndash2744httpsdoiorg10108001431161003743199 2011a

Trigo I F Dacamara C C Viterbo P Roujean J Olesen FBarroso C Camacho-de-coca F Freitas S C Garciacutea-haro JGeiger B Ghilain N Meliaacute J Pessanha L and ArboledaA LandSAF leaf area index available at httpslandsafipmaptenproductsvegetationlai (last access 15 November 2020)2011b

Venturas M D Sperry J S and Hacke U G Plantxylem hydraulics What we understand current researchand future challenges J Integr Plant Biol 59 356ndash389httpsdoiorg101111jipb12534 2017

Ward E J Bell D M Clark J S and Oren R Hydraulictime constants for transpiration of loblolly pine at a free-aircarbon dioxide enrichment site Tree Physiol 33 123ndash134httpsdoiorg101093treephystps114 2013

West A G Hultine K R Jackson T L and EhleringerJ R Differential summer water use by Pinus edulisand Juniperus osteosperma reflects contrasting hy-draulic characteristics Tree Physiol 27 1711ndash1720httpsdoiorg101093treephys27121711 2007

Wigneron J P Jackson T J OrsquoNeill P De Lannoy G deRosnay P Walker J P Ferrazzoli P Mironov V BircherS Grant J P Kurum M Schwank M Munoz-Sabater JDas N Royer A Al-Yaari A Al Bitar A Fernandez-Moran R Lawrence H Mialon A Parrens M RichaumeP Delwart S and Kerr Y Modelling the passive mi-crowave signature from land surfaces A review of recent re-sults and application to the L-band SMOS amp SMAP soil mois-ture retrieval algorithms Remote Sens Environ 192 238ndash262httpsdoiorg101016jrse201701024 2017

Xu X Medvigy D Powers J S Becknell J M andGuan K Diversity in plant hydraulic traits explains sea-sonal and inter-annual variations of vegetation dynamics inseasonally dry tropical forests New Phytol 212 80ndash95httpsdoiorg101111nph14009 2016

Zhang Y Zhou S Gentine P and Xiao X Can vegetation opti-cal depth reflect changes in leaf water potential during soil mois-ture dry-down events Remote Sens Environ 234 111451httpsdoiorg101016jrse2019111451 2019

Zhuang J Yu G-R and Nakayama K A Series RCL CircuitTheory for Analyzing Non-Steady-State Water Uptake of MaizePlants Sci Rep 4 6720 httpsdoiorg101038srep067202014

Zwieback S Bosch D D Cosh M H Starks P J and Berg AVegetation-soil moisture coupling metrics from dual-polarizationmicrowave radiometry using regularization Remote Sens En-viron 231 111257 httpsdoiorg101016jrse20191112572019

Abstract
Introduction
Methods
- Datasets
- Soil moisture pulse identification
- Vegetation pulse response timescale estimation and analysis
- Satellite plant water content response uncertainty analysis
- Plant hydraulic model simulations
- - Results
  - - Global plant water content characteristic responses and timescales
    - Growth influence on plant water content increase timescales
    - Pulse condition influence on plant water content increase timescales
    - - Discussion
      - Plant water uptake timescale variation across climates
        
        Growth impact on dryland plant water uptake timescales
        
        Slow dryland plant rehydration mechanisms
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        Author contributions
        
        Competing interests
        
        Special issue statement
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 2: Patterns of plant rehydration and growth following pulses ...

Ogle et al 2015 Ogle and Reynolds 2004 Sperry et al2016) Characterizing these timescales and their dependen-cies across biomes will increase our understanding of whole-plant behavior and assist plant hydraulic parameterizations inland surface models to better assess plant water stress (Bonanet al 2014 Fisher et al 2018 Kennedy et al 2019 Lin etal 2019 Tai et al 2017 Xu et al 2016) However the fun-damentals of these pulse water use durations especially soil-to-plant water storage timescales remain unknown globally

Soil and plant water content measurements can be used tocharacterize the soilndashplant hydraulic system and understandplant water storage timescales but they are laborious and arethus often constrained to a single location Alternatively mi-crowave remote sensing satellites provide plant water contentobservations across the globe at near-daily sampling frequen-cies (Entekhabi et al 2010 Kerr et al 2010 Konings et al2016) Although these observations are at a coarse resolution(tens of kilometers) their ecosystem-scale resolution and re-lationship to leaf water potential (Momen et al 2017 Zhanget al 2019) make them a useful tool for studies of ecosystemplantndashwater relations (Feldman et al 2018 Konings et al2019 Konings and Gentine 2017) However the plant watercontent measurements are a function of both relative watercontent and dry biomass thus making them additionally sen-sitive to biomass and growth (Momen et al 2017 Zhanget al 2019) This is nevertheless an advantage because theplant water content sensitivity to both water potential and drybiomass allows evaluation of timescales of multiple whole-plant mechanisms at the landscape scale

With regard to plant rehydration timescales plant wateruptake and storage timescales have typically been assessedusing an electric circuit analogy with the timescale of in-terest being the plant resistance times capacitance or RCtime constant (Phillips et al 1997 2004 Ward et al 2013)The RC time constant quantifies the time required for leafor xylem water potential to reach 63 of its equilibriumvalue following a soil moisture or transpiration perturbationMeasured RC time constants vary from minutes for grassesto hours for trees (Hunt and Nobel 1987 Nobel and Jor-dan 1983 Phillips et al 1997 2004 Ward et al 2013)According to this theory plant rehydration after rain pulsesshould occur within a day By contrast field pulse experi-ments of grass and shrub species primarily in dryland envi-ronments (broadly annual rainfall less than 500 mm) com-monly show multi-day predawn water potential increasesafter rewetting pulses (Fravolini et al 2005 Huxman etal 2004 Ignace et al 2007 West et al 2007) Multi-daypredawn water potential increases would lengthen plant wa-ter content timescales These multi-day water potential in-creases appear to be related to recovery from water limita-tion between storms which highlights the potential impactof antecedent moisture conditions on plant responses (Guoand Ogle 2019 Ogle et al 2015 Plaut et al 2013) Hy-draulic limitations from previously dry conditions have alsobeen observed driving multi-day recovery of leaf gas ex-

change after soil rewetting (Blackman et al 2009 Brodribband Cochard 2009 Chen et al 2009 Huxman et al 2004Martorell et al 2014) These multi-day hydraulic responseobservations call into question whether hydraulic responsetimescales are consistently sub-daily across global biomesThey also highlight the unknown role of moisture pulse char-acteristics (antecedent soil conditions and pulse magnitude)on these timescales

With regard to timescales of growth while growth isknown to occur on seasonal timescales there is evidencethat growth can occur following rainfall events as under thepulse reserve hypothesis (Noy-Meir 1973) Specifically dry-land measurements suggest that growth can occur over daysto weeks following a pulse (Angert et al 2007 Doughertyet al 1996 Hermance et al 2015 Novoplansky and Gold-berg 2001 Post and Knapp 2019 Sher et al 2004) Alsoecosystem growth responses in drylands have been mod-eled previously with a 1ndash5 d lag (depending on plant type)and a decaying persistence over 1ndash2-week scales (Ogle andReynolds 2004 Reynolds et al 2004) Ultimately pulse-driven growth following rainfall would lengthen plant watercontent timescales by increasing the total plant water storagecapacity

Here we evaluate the duration of total plant water con-tent increases following rainfall pulse events Under nom-inal moisture conditions with no growth one would ex-pect sub-daily plant water content increases based on RCtime constants Slow rehydration andor growth would likelyextend these timescales to multiple days We ask acrossglobal biomes do plant water uptake responses to soilmoisture pulses ever occur beyond a day and what arethese timescales How do pulse characteristics (pulse mag-nitude moisture pre-conditions) and growth influence thesetimescales Do attributes of the moisture pulse (pulse mag-nitude moisture pre-conditions) favor plant growth versusrehydration To address these questions we use microwaveremote sensing of total plant water content a combinationof dry biomass and relative water content following an ap-proach for rain pulse studies originally developed in Feld-man et al (2018) In order to better understand potentialmechanisms underlying remote-sensing-observed timescalevariations we also discuss the observed timescales and theirdrivers in the context of a SPAC model

2 Methods

21 Datasets

We use 4 years (1 April 2015 to 31 March 2019) of soil mois-ture and plant water content observations from the Soil Mois-ture Active Passive (SMAP) satellite (Entekhabi et al 2010)SMAP measures the low-frequency microwave (14 GHz) ra-diation emitted from Earthrsquos surface The radiation signal isin units of temperature or brightness temperature (TB) The

3 Results

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 3: Patterns of plant rehydration and growth following pulses ...

3 Results

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 4: Patterns of plant rehydration and growth following pulses ...

3 Results

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 5: Patterns of plant rehydration and growth following pulses ...

3 Results

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 6: Patterns of plant rehydration and growth following pulses ...

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 7: Patterns of plant rehydration and growth following pulses ...

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 8: Patterns of plant rehydration and growth following pulses ...

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 9: Patterns of plant rehydration and growth following pulses ...

4 Discussion

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 10: Patterns of plant rehydration and growth following pulses ...

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 11: Patterns of plant rehydration and growth following pulses ...

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 12: Patterns of plant rehydration and growth following pulses ...

5 Conclusions

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 13: Patterns of plant rehydration and growth following pulses ...

References

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 14: Patterns of plant rehydration and growth following pulses ...

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 15: Patterns of plant rehydration and growth following pulses ...

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 16: Patterns of plant rehydration and growth following pulses ...

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Page 17: Patterns of plant rehydration and growth following pulses ...

Abstract
Introduction
Methods
- Datasets
- - Results
    - - Discussion
        
        
        
        Conclusions
        
        Code availability
        
        Data availability
        
        Supplement
        
        
        Competing interests
        
        
        Acknowledgements
        
        Financial support
        
        Review statement
        
        References

Date post:	23-Jan-2022
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

Patterns of plant rehydration and growth following pulses ...

Documents