Science and Societal Objectives by mikeholy


									Why a Hydrology Mission Needs Two-Dimensional Acquisitions of Water Surface Elevations
by Nelly M. Mognard and Douglas E. Alsdorf
A fundamental problem in our understanding of the global water cycle is the measurement and prediction of
water flows across floodplains and wetlands. Fresh water bodies cover at least 4% of the earth’s terrestrial
surface whereas tropical wetlands, particularly in the Amazon Basin, occupy nearly 20% of their watershed.
These vast water bodies are significant stores of freshwater that are often unaccounted in climate and water
cycle models. Predictions of flooding discharge and related societal hazards are particularly difficult because the
flows are spatially complex with both vast diffusive and locally confined hydraulics. This complexity leads to a
rich variety of carbon, nutrient, and sediment dynamics within the wetland ecology. However, our ability to model
and hence predict the hydrologic, ecologic, and societal consequences of floods is greatly limited by the
complete lack of water height (h) measurements virtually anywhere on the globe during the passage of any given
flood wave. Using spaceborne interferometric synthetic aperture radar (SAR) measurements, we show that
changes in flood water heights (dh/dt) are far more complex than typically assumed. Typical modeling
approaches assume that floodplain waters are horizontal and equivalent to those measured in the adjacent main
channel yet such assumptions do not match our observed measurements. Instead, we find that during the
passage of a flood wave geomorphic features such as small floodplain channels can act as not only as conduits
of flow, but surprisingly as barriers. Both point-based stream gauge and profiling altimetric methods of measuring
these water surface elevations and their changes are incapable of capturing the inherent dynamics. For
example, using a profiling altimeter and a 16-day orbital repeat cycle, like that of Terra, misses ~30% of the
rivers and ~70% of the lakes in the global data bases. Restricting the study to the largest rivers and lakes
provides better coverage, but significant water bodies are still missed. Furthermore, the rivers which are covered
can have only a few visits per cycle, leading to problems with slope calculations. Instead, a high-resolution,
image-based approach with broad, two-dimensional acquisitions of h, dh/dt, and dh/dx are required to answer
important hydrologic questions. A 120 km wide swath instrument misses very few lakes or rivers: ~1% for 16-day
repeat and ~7% for 10-day repeat. Therefore, an international team is proposing the Water Elevation Recovery
mission (WatER). A key technology of the WatER mission is a Ka-band Radar INterferometer (KaRIN) which is
capable of the required high-resolution 2D measurements.
1.    Introduction
Terrestrial surface water is absolutely essential to life, economies, environment, climate, and weather. Both
national and local economies rely on flowing rivers to transport storm waters, sewage, and other effluents away
from cities as well as offering major shipping lanes to inland areas. The ecologies of wetlands and floodplains
depend on surface water flows to deliver nutrients and to exchange carbon and sediments. Surface waters play
a role in global climate through energy and water mass exchange with the lower atmosphere. Moreover, local
weather is strongly affected by the surface area of nearby water bodies. Runoff is a strong indicator of
accumulated precipitation throughout a watershed and large, periodically flooded wetlands provide vast surfaces
for evaporation as well as water storage. Earth’s six billion people critically rely upon surface water availability
for domestic use, agriculture and industry, while human health is impacted by water borne diseases (e.g.,
disease vectors related to malaria). In fact, national defence issues are related to surface water, particularly via
politically charged water impoundment projects. The global water issues will have large effects on many of the
world’s major decisions in the next decades and will require operational monitoring tools to support water
policies (for ex. the EU Water Framework Directive).
Although improved description of the terrestrial branch of the global water cycle is now recognised as being of
major importance for climate research as well as for inventory and management of water resources and water
related ecosystems, the global distribution and spatio-temporal variations of terrestrial waters are still poorly
known because routine in-situ observations are not available globally or even regionally [Vörösmarty et al.,
2000]. So far, global estimates of spatio-temporal change of land water stored in surface reservoirs and soils
mostly rely on land surface models.
The Water And Terrestrial Elevation Recovery Hydrosphere Mapper (WATER HM) mission addresses this
shortfall by providing synoptic measurements of terrestrial water elevations at the spatial and temporal scales
necessary to address these fundamental questions. Specifically, WATER HM is designed to answer the
question of: Where and how freshwater storage varies at time scales from days to years and spatial scales of
tens of meters to thousands of kilometres.

2.    Difficulties with In-Situ Measurements and existing remote sensing techniques
In-situ gauging networks that provide time series of water levels and discharge rates have been installed for
several decades in many river basins, distributed non-uniformly throughout the world. The in-situ measurements

      Science and Societal Objectives 702d6acf-7cef-4bff-99f6-0d8e66c0fdee.doc
are used for studies of regional climate variability as well as for socio-economic applications (e.g., water
resources allocation, navigation, land use, infrastructures, hydroelectric energy, flood hazards) and
environmental studies (rivers, lakes, wetlands and floodplains hydroecology). In-situ methods are essentially a
one-dimensional, point-based sampling of the water surface that relies on well defined channel boundaries to
confine the flow. Yet water flow and storage changes across wetlands and floodplains are spatially complex with
both vast diffusive flows and narrow confined hydraulics. This complexity is fundamentally a three-dimensional
process varying in space and time, which cannot be adequately sampled with one-dimensional approaches. In
addition, 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 [Stokstad, 1999]. Therefore, our ability to measure, monitor, and forecast
European and global supplies of fresh water using in-situ methods is essentially impossible because of (1) the
decline in the numbers of gauges worldwide [Vörösmarty et al., 2001], (2) the poor economic and infrastructure
problems that exist for non-industrialized nations, and (3) the physics of water flow across vast lowlands.
During the past decade, remote sensing techniques (satellite altimetry, radar and optical imagery, active and
passive microwave techniques, InSAR, space gravimetry) have been used to monitor some components of the
water cycle in large river basins [Cazenave et al., 2004]. Radar altimetry in particular has been used extensively
in the recent years to monitor water levels of lakes, rivers, inland seas, floodplains and wetlands [e.g., Birkett
1995, 1998; Birkett et al., 2002; Mercier et al., 2002, Maheu et al., 2003; Kouraev et al., 2004]. A few examples
of altimetry-derived water level time series over rivers are presented in the Figure 1.
However, 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 complex and multi-peaked due to
interfering reflections from water, vegetation canopy and rough topography. These effects result in less valid
data than over oceans. Systematic reprocessing of raw radar waveforms with optimized algorithms provides
decade-long time series of terrestrial water levels, at least over large (> 1 km width) rivers [Berry, 2003].
Nevertheless, presently operating radar altimeters built to sample the surface of the open ocean miss numerous
water bodies between orbital tracks. Repeat-pass SAR interferometry has been shown to offer important
information about floodplains in measuring small water level changes [Alsdorf et al., 2000]. However, poor
temporal resolutions are associated with repeat-pass interferometric SAR (repeat orbits are usually monthly, at
best, leading to a prolonged ∂t in the ∂h/∂t mapping). Off-nadir single-pass interferometric SAR will not work over
open-water, instead it requires special hydro-geomorphologies of flooded vegetation [Alsdorf et al., 2000; 2001a,
2001b; Lu et al., 2005; Kim et al., 2005].
Optical sensors cannot penetrate the canopy of inundated vegetation and typically fail to image water surfaces

  Fig.1: Water level time series over the Niger (upper panel; left), Yangtse (upper panel; right),
  Indus (lower panel; left) and Danube (lower panel; right) based on Topex/Poseidon (source: LEGOS).
when clouds or smoke are present [e.g., Smith, 1997]. The prevalent vegetation and atmospheric conditions in
the tropics lead to much reduced performances for technologies operating in and near the optical spectrum. By
themselves, none of the presently operating satellite technologies supply the water volume (except the current
GRACE gravimetry mission; but its resolution, on the order of 500 km, is still poor; Tapley et al., 2004; Ramillien
et al., 2005) and hydraulic measurements are needed to accurately model the water cycle and to guide water
management [Alsdorf et al., 2003; Alsdorf and Lettenmaier, 2003]. Although the Shuttle Radar Topography
Mission (SRTM) produced a high spatial resolution image of heights, the errors over water surfaces are quite

      Science and Societal Objectives 702d6acf-7cef-4bff-99f6-0d8e66c0fdee.doc
large and the mission was active for only 11 days sampling period in February 2000, preventing temporal
change studies [e.g., ±5.5 m, LeFavour and Alsdorf, 2005].
3.     Science and Societal Objectives for a surface hydrology mission
The global water cycle consists of continental water budgets, the dynamics of aquatic ecosystems, the
estimation of stream flow, and linkages to human life.
3.1. The Global Water Cycle and climate modelling
Fresh water on land is stored in various reservoirs: snow pack, glaciers, aquifers and other geological
formations, root zone (upper few meters of the soil), and surface waters (rivers, lakes, man-made reservoirs,
wetlands and inundated areas). Land waters are continuously exchanged with the atmosphere and oceans
through energy and water mass fluxes (evaporation, transpiration of the vegetation, surface runoff and
groundwater flow). Many land surface parameters exert a strong influence on the surface fluxes of water and
energy, and thereby on the atmosphere. Although improved description of the terrestrial branch of the global
water cycle is now recognized as being of major importance for climate research as well as for inventory and
control management of water resources, the global distribution and spatio-temporal variations of continental
waters are still poorly known because routine in-situ observations are not available globally.
Over the recent decades, the growing use of climate modelling has given rise to the development of global land
surface models (LSMs) to provide realistic temperature and humidity boundary conditions to atmospheric
models. Unfortunately, the conclusions of these coupled studies rely on the assumption that the land surface
variables are perfectly simulated, which is not easy to verify given the lack of global climatologies for hydrological
variables. It is therefore crucial to develop strategies for validating the LSMs. The main strategy is to use a River
Transport Model to transform the gridded runoff into river discharge that can be compared with available in-situ
measurements. This approach, however, is problematic because of the worldwide decline of in-situ gauges (see
section 1). In addition, many river basins are influenced by human activities (e.g., irrigation) that modify runoff, a
problem for model validation [Vörösmarty and Sahagian, 2000]. Another aspect concerns the interseasonal and
interannual variations in surface water storage volumes as well as their impact on balancing regional differences
between precipitation, evaporation, infiltration and runoff are not well known. Lacking measurements of wetland
locations and sizes, hydrologic models can result in significant mismatches between observed and predicted
discharges at gauge locations. Errors can exceed 100% because wetlands moderate runoff through temporary
storage and provide water surface area for precipitation and evaporation [Coe, 2000].
Although global land system models continue to improve through incorporation of better soils, topography, and
land-use land cover maps, their representations of the surface water balance are still greatly in error, in large
part due to the absence of an adequate observational basis for quantifying river discharge and surface water
storage. Therefore, land surface modellers have very much to gain from current and future satellite observations
to provide global hydrological datasets (water storage, river discharges, etc.) that could be used to evaluate the
models. While the first step is clearly a comparison of simulated and satellite derived physical parameters, the
next step will be data assimilation. For the continental water cycle, the images of h, ∂h/∂x, and ∂h/∂t will be
assimilated into hydrodynamic models to produce estimates of flow discharge. Numerical developments and
increases in computer power mean that such models can now be applied at global basin scales [Coe, 2000;
Bates and Wilson, 2004] at high (1 km or less) spatial resolution and the resulting discharge estimates can be
used, for the first time to estimate the surface runoff component of the terrestrial hydrological cycle. The data will
also contribute to enhanced methods for flood prediction in remote basin and to the validation of hydrodynamic
and flood prediction models worldwide.
These issues are summarized in the following questions, which WATER HM would enable us to address: What
is the spatial and temporal variability in the world’s terrestrial surface water storage? What is the global
distribution of freshwater runoff delivered to the oceans and what is its inter-seasonal and inter-annual
variability? Answering these questions will lead to an improved understanding of water cycle processes and
their interactions with climate.
3.2.   Flow Hydraulics
Floodplains are marked by a rich variety of water sources including overbank flows (regional contributions) as
well as groundwater, hyporheic water, local tributary water, and direct precipitation (local contributions) [Mertes,
1997]. Floodplain flow is equally complex including diffusive transport across broad, flat pans, temporary storage
in lakes of varying morphologies, and slow drainage through a maze of channels of various widths, depths,
degrees of boundary definition, and vegetation densities (Figure 2). This complexity impacts water balance and
wetland ecologies. For example, based on Muskingum modeling, Richey et al. [1989] estimate that the main
Amazon River alone exchanges about 25% of its average annual flow with its adjacent floodplain. Although this
percentage is greater than twice the discharge of the Mississippi River, the estimate is not constrained by any in-
situ floodplain gauges. In fact, Alsdorf [2003] used spaceborne interferometric synthetic aperture radar (SAR)
measurements of the floodplain to demonstrate the possibility of significant errors. Given that the Amazon Basin

       Science and Societal Objectives 702d6acf-7cef-4bff-99f6-0d8e66c0fdee.doc
contains about 750,000 km of annually inundated area [Melack and Forsberg, 2001], the impacts likely extend
                                                          far beyond the mainstem. Similarly, the flow of water
                                                          through braided rivers is nearly impossible to measure
                                                          from a singular gauging point because braided rivers
                                                          contain dynamic channels that increase in number,
                                                          widen, and shift in response to changes in discharge.
                                                          Arid, glacierized, and high-latitude basins all typically
                                                          contain braided rivers, yet their geomorphic complexity
                                                          limits in-situ efforts to measure flow variations related to
                                                          the observed retreat of many of the world’s alpine
                                                          glaciers [Meier, 1984; Haeberli et al., 1989, Mognard and
                                                          Josberger, 2002]. Unfortunately, nearly all of the world’s
                                                          wetlands lack in-situ measurements of storage and flow,
    Fig. 2: Inundated floodplain of the Amazon River      while the remoteness and morphology of many Arctic
    showing the complexity of flow hydraulics.            braided rivers limits gauging methods. Even in developed
    Singular gauges are incapable of measuring the flow   countries, in-situ gauges are widely spaced (10-100km)
    conditions and related storage changes implied by     and become increasingly unreliable at higher flows. Such
    this photo. The ideal solution is a spatial           data cannot be used to understand the development of
    measurement of water heights from a remote            flooding at basin scales or validate flood extent
    platform. (photo courtesy L. Hess)                    predictions from distributed hydraulic models [Bates,
                                                          2004]. Our knowledge of the spatial dynamics of
floodplain inundation is therefore poor and our ability to construct adequate flood prediction models may be
severely constrained. Lacking observations, we cannot answer key scientific questions such as: How much
water is stored on a floodplain and subsequently exchanged with its main channel? What are the spatial
dynamics of floodplain inundation? What are the local and continental-scale responses of braided rivers
to climate induced changes in glacier mass-balances?
3.3.   Lakes, reservoirs and wetlands
In addition to rivers, lakes, reservoirs and wetlands are vital to society because they provide drinking water
supply, hydropower, irrigation, aquaculture, and flood control. Wetlands are important ecological sites that are
protected under various conventions, such as the Convention on Wetlands [Carp, 1972]. However, 40% of the
worlds largest lakes have not been studied and their volumes are therefore approximated [Shiklomanov,
forthcoming]. Hydrologic fluctuations in the world's hundreds of thousands of smaller lakes and wetlands are
virtually unknown. WATER HM will provide scientists with repetitive, centimetric accurate water level
measurements, and this will greatly improve the understanding of storage dynamics of all lakes, reservoirs and
wetlands, globally and locally.
Knowledge of natural variability and long-term changes in lake area and water levels are also vital to scientific
understanding of climate change, emissions to the atmosphere, surface energy budgets and ecology. For
example, satellite imaging of hydrologic fluctuations in nearly 10,000 Siberian lakes and wetlands has recently
identified a previously unrecognized process of lake drainage triggered by climate warming and permafrost thaw
currently underway in the Arctic [Smith et al., 2005]. They also release large quantities of water vapour, carbon
dioxide, and methane to the atmosphere, absorb more solar energy than surrounding terrain, and can export
large quantities of dissolved organic carbon from the land surface to oceans [Frey and Smith, 2005]. Globally,
small lakes and wetlands support a diversity of endemic ecosystems and are vital feeding and nesting sites for a
great number of bird species that migrate over long distances [Alerstam et al., 2001], and support diverse
populations of fish, birds, animals and humans.

       Science and Societal Objectives 702d6acf-7cef-4bff-99f6-0d8e66c0fdee.doc

To top