Modéle de document par défaut CNES version 1.5 mars 1999



Terrestrial water-storage contributions to sea-level rise and variability

P.C.D. Milly (1) (rapporteur),

A. Cazenave (2), J. Famiglietti (3), V. Gornitz (4), K. Laval (5), D. Lettenmaier (6), D. Sahagian (7), J. Wahr (8), C. Wilson (9) (exept for 1st author, names are in alphabetical order)

(1) GFDL/NOAA, Princeton, NJ, USA (cmilly@)

(2) LEGOS/CNES, Toulouse, France (anny.cazenave@cnes.fr)

(3) University of California, Irvine, CA, USA (jfamigli@uci.edu)

(4) NASA/GISS, New York, NY, USA (vgornitz@giss.)

(5) LMD UPMC UMR 8539, Paris, France (laval@lmd.jussieu.fr)

(6) University of Washington, Seattle, WA, USA (dennisl@u.washington.edu)

(7) CSRC/EOS, University of New Hampshire, Durham, NH, USA (gaim@unh.edu)

(8) University of Colorado, Boulder, USA (john.wahr@Colorado.edu)

(9) University of Texas, Austin, Texas, USA (crwilson@mail.utexas.edu)

1. Introduction

1 Purpose and Scope

A gain or loss of water by the continents generally corresponds to an equal loss or gain of water by the oceans, which in turn affects the global mean sea level. In this review, we summarize current understanding and uncertainties on contemporary continent-ocean water exchanges on time scales ranging from seasonal to centennial. We exclude from consideration the exchanges between the ocean and the ice sheets of Greenland and Antarctica, as well as the exchanges between the ocean and mountain glaciers. These exchanges are considered in other reports of the workshop. However, we do include exchange between the oceans and the subsurface continental cryosphere, which apparently is not treated elsewhere.

2 External Constraints on the Contribution of Terrestrial Water to Present-Day Sea-Level Change

Interannual to decadal change in land water storage is a potentially important contributor to global mean sea-level change. Most recent estimates indicate that sea level has been rising by 1.8 mm/yr during the last 50 years and by 3.1 mm/yr during the 1990s (Church et al, 2004, Holgate and Woodworth, 2004, Cazenave and Nerem, 2004). On these time scales, the two main causes of sea-level rise are thermal expansion of the warming oceans and fresh water mass transfer to the oceans from melting mountain glaciers, ice sheets, and other continental reservoirs.

The contributions of thermal expansion and melting mountain glaciers are now reasonably well estimated both for the period of the past few decades and for the 1990s. These two processes account respectively for 0.4 mm/yr and 0.45 mm/yr of sea level rise for the past 4-5 decades and for 1.5 mm/yr and 0.8 mm/yr for the last decade (Levitus et al., 2005, Ishii et al., 2006, Willis et al., 2004, Lombard et al., 2006, Duygerov and Meier, 2005). For the 1990s, thermal expansion plus mountain glacier melting thus contribute 2.3 mm/y, leaving ~0.8mm/yr to be explained by other contributions, such as change in mass of the Greenland and Antarctica ice sheets plus change in other land water stores. Concerning the ice sheets, remote sensing observations available for the recent years have provided the first direct observations of the mass balance of Greenland and Antarctica. Most recent results suggest that, on average, Greenland and Antarctica contributed respectively ~0.2 +/-0.1 and ~0.2 +/- 0.3 mm/yr during the last decade, even though large uncertainty remains (Thomas et al., 2004, Davis et al., 2005, Krabill et al., 2005, Rignot and Kanagaratnam, 2006). Thus, the amount of observed sea level rise apparently not explained by the sum of thermal expansion plus exchange with ice sheets and mountain glaciers, which must be explained by mass exchange with other continental water stores, appears most likely to fall in the range between +0.1 mm/yr and +0.7 mm/yr during recent years (the 1990s), but could conceivably be negligibly small or even slightly negative.

3 Major Domains of Land Water Storage

Water is stored on land in various types of reservoirs, including surface water, ground water, glacial, ice, snowpack, permafrost, and soil moisture. These have varying distribution, and time scales of variation. Each can be driven by both natural and anthropogenic processes.

It is useful for our purposes to divide the land water stores into those that are distributed extensively in area and those that are concentrated in small areas. We shall refer to these as distributed and concentrated stores, respectively. The relative applicabilities of measurement and modeling tools differ between the distributed and concentrated stores, because of the inherent difference in spatial scales.

The distributed stores are snow pack and subsurface waters. Subsurface waters can be subdivided into the saturated zone (“groundwater”) of aquifers and other geological formations, the root zone (upper one or two meters of the soil, whose water is called “soil water” or “soil moisture”), and, in arid regions, the intervening vadose zone. Concentrated stores are the various forms of surface water. These include rivers, lakes, man-made reservoirs, swamps and intermittently inundated land areas.

4 Major Drivers of Variations in Land Water Storage

Changes in land water storage result from climate variations, from direct human intervention in the water cycle, and from human modification of the physical characteristics of the land surface. Climate variations cause changes in the surface water balance, which can increase or decrease water storage; cool and wet climatic anomalies tend to drive storage upward, and warm and dry anomalies tend to drive storage downward. Major human activities that directly affect storage are the removal of groundwater from storage by pumping in arid regions, the creation of artificial water storage reservoirs by construction of dams on rivers, irrigation of cropland, and the diversion of rivers that recharge closed-basin lakes (i.e., lakes, present only in arid regions, that have no outflow and that lose water only to evaporation). Anthropogenic changes in the physical characteristics of the land surface result from urbanization, agriculture, and forest harvesting (along with forest re-growth).

5 Overview

In this paper, we first review the status of understanding of change in land water storage caused by climate variations and human activities; our main concern is with the effect of such changes on sea level. We attempt to identify the major sources of uncertainty. On the basis of the review, we then provide a list of recommendations, concerning both modeling and observations, with the objective of improving this still poorly constrained contributor to sea level change. The paper is organized into the following sections:

• Analysis tools: measurements and models

• Climate-driven changes of land water storage

• Anthropogenic changes of land water storage

• Synthesis

• Recommendations

2. Analysis Tools

6 In-Situ Observations

In-situ observations have their greatest potential utility for the evaluation of changes in concentrated stores, which we have already identified mainly as surface-water bodies. For any given lake, the storage variation is easily determined by monitoring of lake level and knowledge of lake area. The power of in-situ observations to quantify directly the global storage variations in concentrated stores depends on the distribution of the global storage variation across stores of various sizes and the number distribution of the store sizes. If global storage change is mostly explained by a small number of large stores, then in-situ assessment is practical; if a large part of storage change is contained in a large number of small stores, then analysis by in-situ measurements alone is impractical, but some combination of in-situ measurements and statistical extrapolation may be informative.

In-situ observations of concentrated stores also facilitate the evaluation and calibration of remote-sensing methods (i.e., satellite altimetry, discussed below).

Because of the heterogeneity of distributed stores and the inherently small spatial sampling scale of in-situ observations, such observations alone are not of great direct utility for estimating climate-driven changes of distributed stores. Their potential value lies more in their usefulness for evaluation and calibration of models that can be used to generate global storage estimates. Additionally, in-situ measurements might be of substantial benefit in the assessment of concentrated changes in distributed stores; such changes arise where anthropogenic disturbance is concentrated in nature, as in the case of withdrawal of groundwater by pumping in arid regions.

7 Satellite Observations

As in the case of in-situ observations, observations from satellite platforms serve the dual purpose of directly yielding estimates of storage and of supporting the development of models that can provide less direct estimates (e.g., Alsdorf et al., 2003, Alsdorf and Lettenmaier, 2003, Cazenave et al., 2004). We focus here on two types of space-based systems of most direct current relevance for observation of land water storage: gravimetric and altimetric systems.

1 GRACE space gravity data

The recently (2002) launched GRACE space mission now provides an invaluable set of new observations allowing us to quantify the spatio-temporal change of the total land water storage (underground and surface waters, snow and ice mass changes). In addition the GRACE data over the oceanic domain can provide information regarding the ocean mass change (hence one of the two contributions to sea level change, i.e., that resulting from water mass addition due to land ice melt and exchange with terrestrial storage). Moreover, GRACE has the capability to measure the combined effects of climatic plus anthropogenic change in land water storage.

In March 2002, a new generation of gravity missions was launched: the Gravity Recovery and Climate Experiment (GRACE) space mission (Tapley et al., 2004 a,b). The objective of GRACE is to measure spatio-temporal variations of the gravity field with an unprecedented resolution and precision, over time scales ranging from a few months to several years. As gravity is an integral of mass, these spatio-temporal gravity variations represent horizontal mass redistributions only to the extent they are assumed to be caused by surface water changes. On time scales from months to decades, mass redistribution mainly occurs inside the surface fluid envelopes (oceans, atmosphere, ice caps, continental reservoirs) and is related to climate variability (both from natural and anthropogenic sources). The main application of GRACE is quantifying the terrestrial hydrological cycle through measurements of vertically-integrated water mass changes inside aquifers, soil, surface reservoirs and snow pack, with a precision of a few cm in terms of water height and a spatial resolution of ∼500 km (e.g., Wahr et al., 1998, Rodell and Famiglietti, 1999, Wahr et al., 2004, Seo et al., 2005, Ramillien et al., 2005a, Schmidt et al., 2006, Chen et al., 2004, 2005a,b, Swenson and Milly, 2006, Ngo-Duc et al., 2006). From these quantities, it is also possible to estimate some hydrological parameters such as precipitation minus evapotranspiration, evapotranspitation and total basin discharge (e.g., Rodell et al., 2005, Wahr et al., 2006, Syed et al., 2006). Besides, GRACE is also able to measure the mass balance of the ice sheets and corresponding contribution to sea level (Velicogna and Wahr, 2005, 2006, Ramillien et al., 2006), ocean mass change (Chambers et al., 2004), as well as geographically averaged thermal expansion when combined with satellite altimetry (Garcia et al., 2006, Chambers, 2006).

The data set provided to GRACE users by the GRACE project consists of monthly sets of spherical harmonic geoid coefficients (and associated uncertainties), up to degree and order 100, since April 2002. These coefficients derived from raw tracking measurements (GRACE consists of a pair of satellites whose distance, absolute positions and velocities are continuously monitored) are currently computed by several groups: the Center for Space Research (CSR) and Jet Propulsion Laboratory (JPL) in the USA and the GeoForschungsZentrum (GFZ) in Germany. Recently another group is providing GRACE geoid solutions (the GRGS Group in Toulouse).

The static component of the gravity field corresponds to nearly 99 per cent of the total field, mainly due to solid Earth contributions. Time variable geoid anomalies are in general expressed with respect to the static field, considering that the latter can be approximated by the temporal mean of a long enough series of GRACE monthly geoids (typically several years).

The geoid solutions are provided as a set of so called ‘spherical harmonic’ coefficients, {δCnm(t) and δSnm(t) }, where n and m are the degree and order respectively, such that the time-variable geoid writes:

[pic] (1)

where N is the maximum degree of the decomposition, θ is the co-latitude, λ is the longitude and Pnm is the associated Legendre polynomial. Degree N is equivalent to a spatial wavelength λ, with λ ~ 2πRe/N (Re is mean Earth’s radius). The time variable geoid is classically expressed as a surface load function δq(θ, λ, t) related to the geoid spherical harmonics through (e.g., Wahr et al., 1998):

[pic] (2)

where k’n represents the Love numbers that allows taking into account the elastic compensation of the Earth to surface load. M is Earth’s mass. Because of orthogonality of Legendre polynomials, the spherical harmonic coefficients of the GRACE geoids of a given degree and order are linearly related to corresponding spherical harmonic coefficients of the load function q, so that it is easy to deduce this surface load from (3). The load is further simply expressed in terms of equivalent water height, either globally or regionally, and then usable for further comparison with LSMs outputs and other applications.

The very first studies have validated the capability of GRACE to recover vertically integrated water mass change (sum of underground, soil and surface waters, and where appropriate snow) in large river basins, by comparing GRACE results with global LSMs predictions (e.g., Tapley et al., 2004b, Wahr et al., 2004, Chen et al., 2005b, 2005, Ramillien et al., 2005, etc.).

Recently Wahr et al. (2006) estimated the accuracy of GRACE water mass determinations. They showed that the error of individual months GRACE solutions depends on latitude only, and is on the order of 8 mm (equivalent water height, ewh) near the pole and ~25 mm ewh at low latitudes, for 750 km resolution.

Fig.1 shows a map of total land water storage change (trends) from GRACE over a 3-year period (mid-2002 to mid-2005). The signature of large hydrographic basins is clearly visible.

In some recent studies, it has been shown that GRACE is quite helpful for improving LSMs and validate model predictions (e.g., Swenson and Milly, 2006, Ngo-Duc et al., 2006) (see section 2.3).

Other studies have focused on different aspects of sea level change using GRACE. For example, Chen et al. (2005b) have estimated the contribution of total land water change (based on GRACE) to the seasonal mean sea level. Accounting for the small water vapor effect and correcting the altimetry-based annual mean sea level for thermal expansion, they found very good agreement between GRACE-based land water storage and non-steric global mean sea level (Fig. 2). Analyzing GRACE data over the oceanic domain only, Chambers et al. (2004) were able to directly determine to total water mass contribution to sea level. Here also a very good agreement was found with the non steric global mean sea level. These results are very encouraging as they indicate that when the GRACE time series lengthens, it will be possible to measure the interannual contribution of land waters (climate-driven plus anthropogenic) either by computing the terrestrial water storage change or by estimating directly the ocean mass change component. In addition, by combining GRACE-based ocean mass change component with satellite altimetry-based global mean sea level, it is possible to estimate thermal expansion, without resorting to in situ hydrographic measurements (e.g., Chambers et al., 2006, Garcia et al., 2006) (see the ‘position paper’ of the ‘Thermal Expansion’ session). Finally other studies have proposed preliminary estimates of the ice sheets mass balance using GRACE and associated contribution to sea level change (Velicogna and Wahr, 2005, 2006; Ramillien et al., 2006).

2 Satellite Altimetry

During the past decade, satellite radar altimetry has much been used to monitor water levels of inland seas, lakes, floodplains and wetlands (e.g., Birkett 1998; Birkett et al., 2002; Mercier et al., 2002, Maheu et al., 2003; Berry et al., 2005, Frappart et al., 2005). Although conventional nadir viewing altimetry has a number of limitations over land because radar waveforms (e.g., raw radar altimetry echoes after reflection on the land surface) are more complex than their oceanic counterparts due to interfering reflections from water, vegetation canopy and rough topography, this technique has proved quite useful to measure surface elevation of surface water bodies. >12-year long water level time series based on the Topex/Poseidon (1992-2006) and Jason-1(2001- ) altimetry missions are now available over nearly one thousand continental lakes and ‘virtual stations’ on rivers (intersection of the satellite ground track with the river), with a typical height precision of 20 cm and temporal resolution of 10-days. With the denser groundtrack coverage of the ERS-2 (1995- ) and ENVISAT (2002- ) altimetry missions, it is several thousands of surface water bodies that can be continuously monitored. But in the latter case, the revisit time is longer (of 35-days). Several data bases have been developed recently, providing direct access through the internet to altimetry-derived surface water level time series (e.g., for lakes, the HYDROWEB data base for lakes, rivers, floodplains and man-made reservoirs, and the ‘River and Lakes’ data base ). In some cases, it is possible to estimate river discharge time series from altimetry-derived water levels (e.g., Kouraev et al., 2004). Moreover, over floodplains, combining altimetry-based water levels with radar or visible satellite imagery allows monitoring surface water volume change, in particular during flooding periods (e.g., Frappart et al., 2005).

In-situ gauging networks providing time series of water levels and discharge rates have been installed for several decades in many river basins, distributed non-uniformly throughout the world. However, gauging stations are scarce or even absent in parts of large river basins due to geographical, political or economic limitations. For example, over 20% of the freshwater discharge to the Arctic Ocean is ungauged and surface water across much of Africa and portions of the Arctic is either not measured or has experienced the loss of over two thirds of the gauges (e.g., Shiklomanov et al., 2002). Given the decline in the numbers of gauges worldwide, the poor economic and infrastructure problems that exist for non-industrialized nations, and the physics of water flow across vast lowlands, space-based measurements of surface waters elevation change (and inferred discharge rates when possible) are of great value for a number of applications in land hydrology. Among these, hydrological models validation using altimetry-derived river discharges and surface water volume changes, and direct estimates of man-made reservoirs volume change through time, are of direct interest for sea level studies.

8 Models of Water Storage

Land waters are continuously exchanged with the atmosphere and oceans through water mass fluxes (evaporation, transpiration of the vegetation, surface runoff and groundwater flow). However, the global distribution and spatio-temporal variations of continental water fluxes and stores are poorly known, because comprehensive observations are not available globally. Land surface models (LSMs) provide a link between water storage and variables that are observed or derived from data. LSMs compute the water and energy balance at the earth surface, yielding time variations of water storage in response to prescribed variations of near-surface atmospheric data. The required atmospheric data are the near-surface atmospheric state (temperature, humidity and wind) and the incident water and energy fluxes from the atmosphere (precipitation and radiation). These are estimated from syntheses of observational analyses and atmospheric model “reanalyses” when the LSM is driven in “stand-alone” mode. Alternatively, they can be simulated by an atmospheric general circulation model when the LSM is run in “coupled” mode.

Before describing LSMs further, it seems important to note that such models were not designed to perform calculations of water storage on land, but rather to calculate fluxes from land to atmosphere for the purpose of atmospheric modelling. This distinction is important, because a model can do very well calculating fluxes and still make large errors, say, in computation of long-term trends in storage. This is possible because storage is a small term in long-term average water balance of land. Only recently have a small number of LSMs been exercised with the problem of land water storage assessment, and it can be expected that further model developments may be needed as this application is further developed.

Global LSMs vary greatly in degree of physical realism, spatial resolution, and explicit representation of vertical and horizontal variability, and a comprehensive review is beyond the scope of this report. Instead, we present only a brief outline of the characteristics of LSMs. An LSM usually divides the global land mass on a regular longitude-latitude grid, with horizontal resolution anywhere from a fraction of a degree (more common in stand-alone applications) to two or three degrees (in atmospheric-coupled applications). Some LSMs include sub-grid heterogeneity by tracking the state of multiple sub-areas, or tiles, that are all assumed to experience the same atmospheric forcing. A time step on the order of an hour typically is used. For each grid cell or tile, the land is divided vertically into a vegetation layer, a snow pack, and a subsurface (“soil”) domain. One or more of these, most commonly the subsurface domain, may be further discretized vertically or simply separated into a root zone and a groundwater layer. Many-layer models do not explicitly distinguish “soil moisture” and “groundwater,” but are nevertheless capable of generating the unsaturated and saturated zones to which these terms refer.

Dynamic equations are used to describe the fluxes among the various layers. Interception storage on the vegetation is a balance of precipitation, throughfall, and evaporation; evaporation is limited by energy availability, which is also tracked for the various layers. Throughfall of snow forms a snowpack; sublimation and snowmelt (again, determined by energy balance) deplete the snow pack. Snowmelt and throughfall of rain infiltrate the soil surface (or run off horizontally) and moisten the surface layers of the soil. Gravity and capillary forces drive the water downward into the soil. Water is drawn from the soil by plant roots, to re-supply water lost form plant tissue as a result of energy-balance-driven transpiration.

Most models have an impermeable boundary a few meters below the surface. Downward-percolating water eventually reaches this boundary and forms a saturated zone that then grows vertically. To leave the soil column, water must flow horizontally; such lateral flow to the river system generally is parameterized in such a way that it increases as the depth of the saturated zone increases.

In some LSMs, when water leaves the soil column either as surface runoff or as lateral outflow from the soil column, it enters a separate model of the river system. The river model consists of a series of river channels, all of which are linked in a tree-like structure that ends at the ocean or at some point of internal drainage. Flows in the river system are usually parameterized simply in terms of a residence time of water in a link. The river model provides an important point of contact between models and observations, because streamflow is readily measured and is a sensitive indicator of the water balance of large land areas.

LSMs can be tested and calibrated in various ways, but generally the available measurements of the extremely heterogeneous fields of snow pack, subsurface water and evaporative fluxes fall far short of what is needed for exhaustive model testing. Models can be tested on a local scale at heavily instrumented sites (e.g., Henderson-Sellers et al., 1995; Chen et al., 1997). Such tests can be useful in identifying major shortcomings in model structure, but can too easily become tuning exercises in which the number of available model parameters exceeds the power of the data to falsify the model. Further, the conclusions of local tests do not easily transfer to the larger spatial scales that are relevant for sea-level assessment.

A complement to local testing of models is use of large river basins as a control volume. Such a practice at least allows accurate determination of the areal average of the runoff flux, by means of conventional streamflow monitoring at a single site. This approach has been taken in the Global Soil Wetness Project (Dirmeyer et al., 1999). The serious shortcoming of this approach is that the basin is treated as a black box; an adequate simulation of streamflow does not ensure a realistic simulation of storage change within the basin.

The local and river-basin approaches to model evaluation mentioned above are both normally implemented in a “stand-alone” model. Such a framework can easily lead to incorrect conclusions if the input atmospheric forcing is not carefully evaluated and adjusted for systematic bias (Milly, 1994).

GRACE is now allowing evaluation of temporal variation in continental-scale storage computed in LSMs. A number of investigators (e.g., Wahr et al., 2004, Ramillien et al., 2005, Ellett et al., 2005, Seo et al., 2006, Chen et al., 2005, etc.) made preliminary comparisons of GRACE water storage estimates with estimates from stand-alone LSM simulations. Swenson and Milly (2006) examined land water storage variations in several climate models that use LSMs to describe land processes. They found substantial model-specific biases in both amplitude and phase of annual storage variations, particularly in low latitudes, and suggested that these were partially associated with sub-optimal descriptions of storage in the models. Ngo-Duc et al. (2006) show striking improvement in the agreement between simulated and GRACE-observed seasonal variations of water storage when a river model that has been calibrated on streamflow measurements is added to the ORCHIDEE LSM.

Because LSMs operate on horizontal scales of tens or hundreds of kilometers, they cannot be readily applied to the smaller-scale problems of anthropogenic disturbance of the hydrosphere, such as those associated with adjustments of the water table as a response to dams. And because LSMs treat only the few meters nearest the land surface, they cannot be applied to examine storage effects associated with groundwater mining and irrigation of arid lands. Of course, because LSMs neglect such processes, care should be exercised in the selection of river basins for LSM evaluation to ensure that anthropogenic processes not cloud the model evaluation.

9 Budget Methods

Storage can be inferred indirectly by application of budget methods. From a global perspective, we can estimate land water storage variations as the opposite of the sum of all other water storage variations in the earth system. This approach was illustrated for the case of decadal trends in Section 1.2. Admittedly, this approach is not useful when the objective is to infer some other (non-land) store as a residual.

Another budget approach takes the land (or some component thereof) as a control volume and assesses storage change as the balance of fluxes to and from the control volume. For example, the change in storage over the Mississippi River basin can, in principle, be estimated as the cumulative difference between precipitation and evaporation from the basin, minus the amount of water leaving the basin as streamflow. The change in storage in a deep aquifer that receives little recharge can be estimated as the cumulative amount of water withdrawn by pumping.

Climate-Driven Changes of Land Water storage

1 Introduction

Because surface waters are in dynamic equilibrium with the atmosphere, the continental water storage budget is climatically controlled. This control is exerted across a range of time scales from seasonal to millions of years. We will focus on the shorter end of that range (seasonal to multidecadal). These are the time scales at which climate fluctuations lead to changes on the same order of magnitude as modern anthropogenic alterations of water storage. These time scales are also those of interest for present-day sea level change.

As is clear from discussions elsewhere in this report, global models are best suited for analysis of climate-driven (as opposed to anthropogenic) changes in continental water storage because of the broad length scales and coarse spatial resolution involved.

A more subtle problem with models deserves mention. As noted in Section 2.3, the ability of LSMs to characterize trends in storage has not been demonstrated and has reason to be doubted. The use of a shallow soil domain for computation may reproduce response correctly on short time scales, but the credibility of its fidelity to the physics becomes increasingly suspect on the longer time scales relevant, say, for interpretation of inter-decadal trends in sea level.

2 Distributed Stores

The storage variations of the distributed stores, from seasonal to interannual and decadal time scales, have been the focus of a series of modelling studies in recent years.

1 Seasonal variation and contribution to sea level

During the past decade, several studies have estimated the land water contribution to the cycle of mean sea level by use of global LSMs (Chen et al., 1998, Minster et al., 1999, Cazenave et al., 2000, Milly et al., 2003, Ngo-Duc et al., 2005a, Chen at al., 2005b, Chambers et al., 2004).The general approach of these studies is to estimate the annual ocean mass component from the satellite altimetry-based global mean sea level, after correcting the latter for the steric component (essentially thermal expansion) and taking into account the small annual variation of atmospheric water vapour, and then to compare the ocean mass component to land water storage based on global LSMs or on GRACE. The annual cycle of global mean sea level has an amplitude of 5 mm, with a maximum in October. Because the annual cycle of steric sea level has about the same amplitude but is in phase opposition, once corrected for steric effects (using climatologies in general), the residual sea level displays an amplitude of 10 mm with a maximum in September. The above studies showed that the annual cycle of sea level can be satisfactorily explained by the annual variation in total land water storage, the snow component being the largest contribution (i.e., 70% of the observed signal is due to the seasonal change in the snow pack).

2 Interannual variation and contribution to sea level

3

The decade-long satellite altimetry time series (available since 1993) now provides information also on the year to year change of the global mean sea-level seasonal amplitude. Particularly strong was this change between 1997 and 1998 because of the El Nino event which occured in 1997.

LSMs can be used also to estimate these interannual changes and to diagnose the source of the variability. Ngo-Duc et al (2005a) computed the seasonal change of global sea level, using the Orchidee LSM. They were able to simulate the drastic contrast in the annual sea level observed between 1997 and 1998. The analysis of the model results showed that the change was caused by the ENSO-driven difference in tropical precipitation over land between these two consecutive years.

4 Interannual to multi decadal variation and contribution to sea level

5

The Land Dynamics (LaD) model developed by Milly and Shmakin (2002) was used by Milly et al. (2003) to quantify the contributions of time-varying storage of terrestrial waters to sea level rise in response to climate change on interannual to decadal time scales. A small positive sea-level trend of 0.12 mm/yr was estimated for the period 1981-2000. This corresponds to an overall decrease in the amount of water mass stored on land. While the long-term trend was very small, large interannual/decadal fluctuations dominated the signal. Subsurface water was the major contributor on interannual time scales.

Recently, Ngo-Duc et al (2005b) ran the LSM Orchidee to assess the total climate-driven land water change, and associated sea level change, for the past 5 decades (Fig.3). No significant trend in sea level due to land waters was visible, but large decadal oscillations produced an overall storage range equivalent to 6 mm sea level. A strong decreasing contribution to sea level was found during the 1970s, followed by a slow increase during the next 20 years; during the period of 1975-1993, Orchidee simulation showed an increase of 0.32 mm/yr. During the common simulated period 1981-1998, the ORCHIDEE and LaD models simulated sea level contributions of 0.08 and 0.12 mm/y respectively. For the 1990s, however, the ORCHIDEE-implied trend in sea level was negative, at about –0.1 mm/yr. As in Milly et al. (2003), the ORCHIDEE variations could be attributed to subsurface water and were caused by precipitation variations, especially in the tropical belt.

Another finding of Ngo-Duc et al (2005b) was a strong anticorrelation (-0.9) between decadal change in the contribution of land water storage to sea level and thermosteric sea level estimated from in-situ hydrographic ocean temperature data (Fig. 3). The implications of this result are twofold: on the decadal time scale, land water storage change partially compensates the effect of thermal expansion on sea level; additionally, ocean heat content exerts an influence on the global water cycle on this decadal time scale. The fact that the land water storage decadal variations (and ocean heat content change as well) are occurring over the tropical belt can be interpreted as the result of the interaction between ocean surface temperature and precipitation over these regions.

3 Major Closed-Basin Lakes

There are numerous large basins on the continents that are not presently filled with water due to current climatic and orographic conditions. There are several other internally draining basins that can store significant volumes of water. These include the Caspian-Aral basin, the various Iranian basins, the Chad basin, and many others. Of these, the Caspian Sea presently holds the greatest volume of water. It is fed primarily by the Volga River, which drains much of European Russia. The level of the Caspian Sea fell about 3 m from 1900 to 1977, with a drop of about 1 m in just a few years during the 1930s. The level then rose more than 2 m into the 1990s. However, net change of the Caspian sea level during the 1990s was small. Fig.4 shows the ‘average’ water level change of the Caspian Sea for 1992-2006, measured by satellite altimetry (combining data from several satellites).

The Aral "Sea" is a related internally draining lake in the same basin as the Caspian, and the volume of water stored above ground has been drastically reduced by diversion (for irrigation of the Karakum desert) of the Amu Darya and Syr Daria, the two rivers that feed the lake. The enhanced evaporation due to irrigation has caused most of the water formerly stored in the lake to leave the basin, and presumably rain elsewhere outside the internally draining basin so it ultimately enters the ocean.

4 Permafrost

In sufficiently cold regions, subsurface water deeper than about a meter remains frozen through the year and is called permafrost. When permafrost thaws, the total amount of water stored in the soil column generally decreases. Indeed, in some regions, the soil contains lenses of almost pure ice whose disappearance explains the irregular changes observed to some landscapes upon thaw. Temperature trends in regions of permafrost generally have been positive in recent decades, and evidence suggests that large-scale thawing of permafrost is underway, perhaps with implications also for water storage (Lawrence and Slater, 2005). As the soil column thaws and drains, the subsurface hydraulic connectivity may be enhanced, potentially leading to more free drainage of the landscape. Recently documented large-scale disappearance of lakes in the zone of discontinuous permafrost (Smith et al., 2005) is evidence of such landscape thaw and drainage. Order-of-magnitude estimates suggest that this phenomenon has potential to be an important contributor to sea level rise in recent years. Unfortunately, cryospheric processes are not well described in LSMs. Clearly this is an area for further research in the immediate future.

Anthropogenic changes of land water storage

1 Artificial Reservoirs

On the basis of recent literature, Gornitz (2001) estimated that the volume impounded behind the world’s largest dams grew by about 5000 km3 during the 20th century. Other estimates are higher (Vörösmarty et al., 1997; Nilsson et al., 2005; Shiklomanov and Rodda, 2003), and the actual value is uncertain because of non-reporting or under-reporting for some countries, and because records generally are not available for the countless reservoirs of smaller capacity (Sahagian, 2000). Here we adopt a value of 7000 km3, which is within the range of published estimates. Most reservoir water was impounded during the second half of the century (Chao, 1995), when the associated average rate of sea-level change was, therefore, about –0.4 mm/yr.

The temporal distribution of reservoir filling is important for interpreting interdecadal changes in the rate of sea-level rise. The temporal distribution of this impoundment reflected in Chao’s (1995) Fig. 2 (which included a large portion, but not all, of the total capacity) implies a slow deceleration in the rate of impoundment. This means that the rate of growth of reservoir storage remained positive throughout the second half of the last century, but the magnitude of the rate declined after the late 1970s. Data provided by Chao (1995) and by Shiklomanov and Rodda (2003) suggest a halving of the rate of growth of total capacity from 1950-1978 to 1978-2000. Additionally, capture of sediment by reservoirs effectively reduces the overall rate of increase in global impoundment volume. For the globe, Gornitz (2001) estimates a storage-capacity decay rate of 1% per year. Taken together, these results suggest that the global effect of impoundments was greater (in absolute value) than –0.4 mm/yr sea-level equivalent before 1978 and smaller than that after 1978. For a halving of the rate in 1978, the pre-1978 rate would be about –0.5 mm/yr and the post-1978 rate would be about –0.25 mm/yr. We therefore adopt a rate of –0.25 mm/yr to characterize recent years. The apparent deceleration in impoundment rate would have contributed in small part to the acceleration of sea-level rise that was observed late in the 20th century. The impoundment history of the largest 100 artificial reservoirs is indicated in Fig.5 (Sahagian , 2000).

When a reservoir fills behind a reservoir, the increase in water depth induces seepage into the subsurface. Over time the rate of seepage will decrease and subsurface storage of water will increase until a new equilibrium is reached. At equilibrium, the change in subsurface storage will cease and the rate of seepage will be balanced by increased evaporation to the atmosphere from the region around the reservoir and increased discharge of groundwater to the river downstream of the dam. The ultimate volume of the change in subsurface storage and its time scale of equilibration, which together will determine the effect on sea level, must depend strongly on local climate, geological framework, and materials used in dam construction. Gornitz (2001) made an estimate of the effect of reservoirs on global groundwater storage. In effect, her estimate combined two assumptions that maximize the groundwater storage effect: infinite equilibration time and no flux from groundwater to streamflow or evaporation. Gornitz also assumed that reservoirs lose 5% of their stored water to groundwater, year upon year, with no stabilization; the resulting effect on sea level was about –0.7 mm/yr. Such a large number implies that dams induce more subsurface storage than surface storage. It would be remarkable if this were true, given that reservoir operators do not consider such additional storage. In view of the apparent absence of hydrogeological observations to support the non-equilibrium assumption on a worldwide basis, we have no reason to reject the null hypothesis that groundwater storage change near reservoirs has a negligible effect on global sea level.

2 Groundwater Mining

The artificial withdrawal of water from the ground by pumping (or by free flow of artesian wells) causes a reduction in storage of water. The reduction in storage causes a reduction in water pressure and induces an adjustment of natural flows. In humid regions, precipitation exceeds the evaporative demand, the voids of the earth fill almost to the land surface with water, and the ground leaks and spills excess water into the river system as runoff. Thus, the water table (the top of the saturated zone) is generally not far from the surface, and the groundwater system is tightly coupled to the other near-surface stores. As a result, groundwater storage rises and falls in response to the seasonal cycle of climate, and even to weather, and removal of water by pumping is eventually compensated by adjustments in the natural water fluxes. Relatively small adjustments in groundwater storage lead to new dynamic equilibria. Nevertheless, in large urban areas of groundwater development in humid regions, large-scale “cones of depression” of water storage do develop. Relevant data are available on a piecemeal basis, but such data have not been systematically analyzed and extrapolated to global scale.

In contrast, arid regions are those where precipitation does not meet the evaporative demand. As a consequence, the soil is desiccated by the atmosphere, and water from precipitation rarely penetrates the ground beyond the reach of atmospheric demand. Such systems can be in disequilibrium for thousands of years, as water that had been delivered to the ground during a wetter climate is gradually transported upward from the lowering water table by increasingly small water-potential gradients. In some cases, aquicludes such as shale layers can "protect" fossil ground water from evaporation. In such an environment, artificial withdrawal of water by pumping leads directly to a progressive decline in water storage until the withdrawal stops, for example, because the store has been depleted. Removal of water in excess of recharge is termed mining.

As a crude approximation, then, we can consider that the amount of groundwater mining can be estimated by the rate of pumping of groundwater over the arid regions of the world. To evaluate the net effect of mining on sea level, we must consider the fate of mined water, and this fate is addressed in the next subsection; here we estimate only the gross, direct effect of the mining (Sahagian et al., 1993, Sahagian, 2000). Gornitz (2001) compiled estimates of mining rates for specific countries from various sources; those explicitly reported rates totaled about 61 km3/yr (or 0.17 mm/yr sea-level rise) both for recent years and for the last half-century.

Gornitz extrapolated the value that was based on reports for specific countries by assuming that the ratio of mining to total groundwater withdrawal was similar globally to what it was in the studied regions. Depending on the details of the extrapolation, this approach led to a wide range of estimates of 0.17-0.77 mm/yr for the gross effect of groundwater mining on sea-level rise. However, groundwater resources are generally renewable in humid regions. Furthermore, according to Shiklomanov (1997, Fig. 4.8), the major groundwater mining operations in the world are found in arid parts of the USA, Australia, and China, and in Mexico, Spain, Algeria, Tunisia, Libya, Egypt, and Saudi Arabia. This list of mining operations coincides closely, though not exactly, with the mining centers explicitly listed in the compilation by Gornitz (2001), suggesting that global extrapolation should not increase the gross effect of mining far above 0.17 mm/yr. Allowing for the exclusion of some regions from Gornitz’s compilation and for the likelihood that mining may be accelerating, so that past literature underestimates its magnitude, we adopt a range of 0.2-0.3 mm/yr sea-level rise for recent years, while acknowledging considerable uncertainty.

The net effect of groundwater mining on sea level will be less than the gross effect discussed above, because some of the water is returned to the subsurface as a result of irrigation, which is discussed next.

3 Irrigation

Irrigation is the transfer of water from a water source to an agricultural field. The removal of the water from the source may cause a drop in some reservoir, while the application to crops may cause an increase in the subsurface storage below the crops. We shall consider first the effects of withdrawal, then the effects of the irrigation itself. We have already considered the effect of groundwater withdrawals on storage, and it remains to consider the effects of surface-water withdrawals.

In humid regions, extraction of water from surface water can have little net effect on storage, for the same reasons already discussed in connection with groundwater mining. In arid regions, whose major rivers receive their flow from outside the arid region, the effect of withdrawing water from a river that normally flows to the ocean is to reduce channel infiltration within the arid region. This leads to a reduction in stored subsurface water, but because the change in area wetted by a flowing river is much smaller than the change in wetted cropland area, we shall ignore this effect here. For arid-region rivers that instead flow to a closed-basin lake, the effect is very different. This is the case of the Aral Sea, which has lost most of its volume as a result of diversion of the Amu Dar'ya and Syr Dar'ya rivers for irrigation. The volume of the Aral Sea declined at a rate of 2.7x1010 m3/yr from 1960 to 1990 (Micklin, 1992), implying an associated sea-level rise averaging 0.07 mm/y during the same period. Since 1990 the storage change has been significantly smaller than this.

We consider now the effects of water application to cropland on the global disposition of water storage. Water applied to crops in humid regions either evaporates or runs off. The effect of disturbance on storage is small in humid regions, as we have already argued. When crops are irrigated in an arid environment, part of the applied water evaporates or is transpired by the plants, and part drains vertically from the plant root zone to recharge the thick unsaturated zone and, possibly, eventually, the deeper saturated zone. Thus, the global change in storage associated with application of irrigation water to cropland is approximately equal to the product of the rate of irrigation of arid-region croplands and the fraction of irrigation water that drains downward from the crop root zone. We estimate the global rate of irrigation as 69% (Gornitz, 2001) of the global rate of water withdrawals for all uses (about 4000 km3/yr, Shiklomanov, 1997), or about 2800 km3/yr (Shiklomanov, 1997). As an approximation, we shall assume that all irrigation water is applied in arid regions, where the need for irrigation is greatest. We do not have an estimate of the fraction of applied irrigation water that drains below the root zone. A perfectly efficient irrigation system would allow no drainage, though it might eventually result in the accumulation of salts in the root zone. Inefficiencies in irrigation and/or intentional flushing of the root zone imply nonzero drainage. In some irrigated arid lands, downward drainage from the root zone and resultant groundwater recharge from irrigation have been of sufficient magnitude to raise the water table into the root zone, creating problems with water logging and salinization, and driving the deep-drainage fraction to zero. If, as assumed by Gornitz (2001), 5% of irrigation water goes into subsurface storage, we find a resultant sea level decline of about –0.4 mm/y. We are aware of no basis to rule out the 5% fraction, nor do we have evidence to support it. The true amount could be negligibly small, or it could be larger in magnitude than this value.

The atmosphere above irrigated land may be moistened relative to its natural state. However, the associated rate of change in water storage is easily shown to be negligible in terms of sea level.

4 Wetland Drainage

In the USA, wetlands have been drained at an average rate of 2.2x109 m2/yr since 1780 (Mitchell, 1990; Mitsch and Gosselink, 1986). Wetland drainage entails removal of standing water, soil moisture, and water in plants having an order of magnitude of a 1-m depth of water (Whigham et al., 1978). From these figures, we obtain an average rate of global sea level rise of 0.006 mm/yr. Global wetland area, estimated as 8.56x1012 m2 by Mitsch and Gosselink (1986), is much larger than that of the USA, but we have little information on trends in global wetland area (Finlayson and van der Valk, 1995; Sahagian and Melack, 1998). In Europe, about half the original wetlands have been drained for agriculture, and nearly half in the rest of the world, although inventories are very incomplete (Moser et al., 1996). If we assume that the fraction of global and USA wetlands drained are both 50%, and if we spread that drainage over the same 220-year period, then we can infer, again very crudely, a global sea-level rise of about 0.06 mm/yr. If much of the assumed global wetland drainage were additionally assumed to occur over a shorter period of time, then this estimate would be higher.

5 Storage Changes Induced by Land-Cover Change and Land Management

Urbanization potentially exerts a strong impact on hydrology in many ways. Replacement of vegetated areas by impermeable pavements and other structures can lead to increased surface runoff, reduced infiltration, and a lowering of the water table. On the other hand, the removal of vegetation also reduces evaporative loss, and water-delivery infrastructure can enhance recharge, leading to increase in groundwater storage. However, as in the case of other effects considered here, quantitative data allowing assessment of the global effects of urbanization are lacking.

Forests store water in living tissue. When a forest is removed, transpiration is eliminated so that runoff is more favored in the hydrologic budget (Meher-Homji, 1991). Depending on local climate and topography, this could lead to more or less water stored in the soil. In a poorly drained environment with low slopes, the loss of a forest could cause the water table to rise as a result of decreased evapotranspiration. Alternatively, loss of a forest could cause increased surface runoff and a reduction in subsurface water fluxes and storage. As discussed elsewhere, the humid regions that are home to forests generally respond quickly to disturbance with a new equilibrium that does not require large changes in storage. Additionally, vegetation re-growth is the most common sequel to deforestation.

6 Atmospheric Water Mass

Though not formally within the scope of our review, we touch briefly here on the water content of the atmosphere. Evidence from global climate models supports a simple thermodynamic control of changes in atmospheric water content on decadal scales. Water content rises in proportion to the saturation vapor pressure of the near-surface atmosphere, which is governed by the Clausius-Clapeyron equation. Thus, a 1-degree C rise in global mean surface temperature translates to a 7% increase of the atmospheric water content; the latter is about 25 mm water equivalent. The 0.2-degree per decade rise in temperature typical of recent years translates to a 0.035-mm/yr increase in atmospheric water content and a sea level change of about –0.05 mm/yr.

Synthesis

In Section 3 we saw that the natural annual variation of land water storage is a dominant control of the annual cycle of global mean sea level. We also saw that climate-driven fluctuations in storage at interannual to decadal scales lead to swings in sea level on the order of a few mm.

Table 1 is an attempt to summarize the status of understanding of land contributions to decadal sea level rise. Decadal trends on the order of a few tenths of a mm per year can be generated by distributed stores in response to climate variations. During the most recent decade of the 1990s, modeled climate-drive trends in the distributed land stores that can be reasonably well modeled by LSMs probably caused sea level change of about –0.1 mm/yr. Climate-driven and anthropogenic change in the Caspian Sea level was also small during the 1990s.

We have also noted, however, that the LSMs used for assessment of land water storage change may not be realistic, particularly when they are asked to describe climate transients in the subsurface cryosphere or in deep unsaturated zones. One climate model analysis that does consider some subsurface cryospheric factors leads to an estimate on the order of 0.1 mm/yr sea level rise during recent years. Additional cryospheric processes, neglected in that model, could make the effect even larger.

The filling of reservoirs in recent decades probably contributed about –0.25 mm/yr to 1990s sea level change; a recent global deceleration in filling of reservoirs (resulting from decreasing construction rate and sedimentation) can explain a small part of the recent (1990s) acceleration in rate of sea level rise. The warming climate contributes about –0.05 mm/yr by increasing the water content of the atmosphere. Groundwater mining contributes an opposite effect, perhaps 0.2-0.3 mm/yr. While the uncertainties in these estimates, particularly for mining, should not be forgotten, nevertheless, all three are based on hard data and are probably not far from the mark.

Analysts have speculated on the possibility that non-equilibrium seepage from surface-water reservoirs and irrigation, and land-use changes may also contribute substantially to the contemporary mass balance of the ocean. We do not present hard evidence here to falsify suggestions that such processes are capable of offsetting sea level rise by as much as 1 mm/yr, nor do we present evidence to support such speculation. The sea-level observations themselves, adjusted for ocean warming, glacier melting, and ice sheet mass balances, do not support a strong negative contribution from land. Absent a stronger case for these effects, it seems reasonable to assume by default that they can be neglected. However, we must admit our ignorance concerning some crucial factors and allow the possibility that one or more of these processes could be a significant player.

When we consider only those processes in Table 1 in which we place medium to high confidence, we obtain a net negative trend in sea level of –0.1 or –0.2 mm/yr. These values are below the most likely range of +0.1 to +0.7 mm/yr deduced in Section 1.2, but the discrepancy is not great, in view of the uncertainties involved in arriving at both ranges. One or more of the processes assigned low confidence in Table 1 could, of course, resolve the discrepancy, but so could errors in the estimates of the more robustly quantified trends.

| |Section |1990s sea level trend |Unidirectional? |

| | |(mm/yr) | |

|MEDIUM TO HIGH CONFIDENCE | | | |

|climate-driven distributed stores |3.2 |-0.1 |no |

|Caspian Sea |3.3 |~0.0 |no |

|reservoir filling |4.1 |-0.25 |yes, but decreasing |

| | | |magnitude |

|groundwater mining |4.2 |+0.2 to +0.3 |yes |

|Aral Sea |4.3 |~0.0 |no |

|atmospheric water storage |4.6 |-0.05 |yes |

|LOW CONFIDENCE | | | |

|permafrost thaw and drainage |3.4 |>0 |yes |

|deep groundwater recharge from surface-water reservoirs |4.1 | ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download