by Mcebisi Mkhwanazi
For decades, satellite imagery has been used for environmental monitoring, with water management being one of the areas of application. For example, many are familiar with the Landsat images showing Lake Chad in Africa shrinking over the years. Now satellites are used for more than just providing visual images. This paper discusses the new applications of remote sensing in water management, focusing on satellite imagery.
Due to the increase in water scarcity, exacerbated by a changing climate, it is every water resource engineer’s wish to be able to account for every drop of water. The path of water from the moment it falls onto the earth’s surface to when it is ‘lost’ to the atmosphere has been thoroughly studied. Several instruments are currently used to analyze the energy interactions on the earth’s surface that result in water escaping to the atmosphere. Such instruments include lysimeters, Bowen ratio energy balance (BREB) systems, scintillometers, and eddy covariance technique. However, these instruments would need to be placed at every unit area of land in order to obtain an accurate estimate of energy fluxes across earth. Imagine the whole earth’s surface littered with such equipment, say every square kilometer; it would not only be messy, but also expensive and time consuming to install, maintain and monitor.
Water accounting has been a challenge for hydrologists, irrigation engineers, and water resource experts in general. Evapotranspiration, which is the combined process of water loss from the ground through evaporation and from plant leaves through transpiration, is the component of the hydrologic cycle that is usually most difficult to to estimate. Accurate estimation of ET is more significant in arid and semi-arid regions as water is scarce in those areas.
In determining the water balance in a catchment or any control volume, the input component is precipitation, usually in the form of rain, which is easy to measure because water is visible and containable. However, the output component is evapotranspiration, which is difficult to measure because the water leaves the catchment as a vapor that we cannot see and can hardly contain. We therefore use air ‘wetness’ and temperature as indirect estimates (surrogates) of the water we could be losing – and this is what most of the instruments mentioned above measure. When we use satellites, the sensors mounted on them provide us with an additional and special pair of ‘eyes’ that enable us to see things we could not otherwise see with our human eyes; they allow us to see beyond the visible range of the electromagnetic spectrum. Now, from space we can look down on earth and estimate how much water is lost from lakes, man-made reservoirs, irrigated fields, and from natural vegetation, even at short time steps (e.g. hourly).
How this works
Satellite-based ET algorithms are often used with the surface energy balance equation, which is based on the principle of conservation of energy. In this case, the solar energy incident on the earth’s surface will be reflected back, heat the surface and consequently the air (sensible heat flux), be transmitted into the ground (soil heat flux), or be used to evaporate water (latent heat flux). In most, if not all cases, at least two of these fluxes will occur concurrently in varying proportions. The form to which incident solar energy is transformed depends on the surface type and condition. There are several space satellites that could help to analyze and interpret the interaction of solar energy with the earth’s surface. For most of my work on satellite-based models of ET under conditions of advections, I used the Landsat series of satellites (Landsat 5, which is now out of use; Landsat 7; and the Landsat Data Continuity Mission, now referred to as Landsat 8).
Several ET algorithms have been developed to estimate the energy fluxes. These analyze the reflected radiances of the various bands of electromagnetic solar radiation, including the visible portion (green, blue and red bands), near infrared, and the thermal infrared. Upon reaching the earth’s surface, a portion of these bands is reflected back to space and can be detected by sensors mounted on satellites. These satellite-sensed radiances can be converted to land surface characteristics such as surface albedo, vegetation indices, emissivities and surface temperatures, which are then used to calculate the various components of the energy balance, that is the net radiation, sensible heat flux, ground heat flux and latent heat flux (which is estimated as a residual and is an energy equivalent of evaporated water). The Surface Energy Balance Algorithm for Land (SEBAL), developed in the Netherlands, and the Mapping Evapotranspiration with Internalized Calibration (METRIC) developed in the US are some of the models widely used and, from these, several variants have been developed.
Benefits and challenges
The beneficial output of this remote sensing technique is that, at the time of each satellite overpass, we are able to determine the energy exchange, on every 30 by 30 meter area of a water body, or vegetated or bare land, without having to install instruments all over. However, there is still the need for some ground instrumentation to validate and improve the accuracy of models. Also, most of these models need ground weather data at short time steps (at least hourly). Without such data they tend to incur some errors, especially under conditions of advection (i.e. when there is additional energy available for evaporation due to warm and dry air horizontally brought onto the surface of interest; this is a common phenomenon in arid and semi-arid areas). In some areas, especially developing countries that have no well-equipped automatic weather stations and therefore lack weather data at short time steps, there are efforts to develop models that are effective despite these data limitations. Together with my advisor Dr. José Chávez, we are developing a model modified from SEBAL that will achieve an improved accuracy in ET estimation even under conditions of advection and that uses the limited weather data that is available (mostly daily averages). This model will make it possible to use satellite-based ET models even in places where these were previously not effective due to inadequate weather data.
Satellite-based ET models are still complicated to use, as they require some experience and a reasonable understanding of biophysics and meteorological concepts. There is still a lot of research to be done; yet the benefits are already being realized. There are several applications of this technology; irrigation engineers use it to determine actual water needs for agricultural purposes and it can be used for hydrologic budget at basin level. My current interest is in using this technology to study climate change. There is consensus that the rate of water loss from water bodies and irrigated landscapes is increasing due to climate change. Now, with the added ability to determine energy fluxes from satellite images, we can go back through time and determine if actual changes in evaporative demand over the years link to our observations of changes in climate. This gives us an improved level of certainty concerning the impact of climate change on water resources and will also help in refining our projections.
Mcebisi Mkhwanazi is a 2010 fellow of the Fulbright Science & Technology Award, from Swaziland. He completed his PhD in Civil Engineering at Colorado State University, Fort Collins.