-
Notifications
You must be signed in to change notification settings - Fork 1
Home
ELEO remote sensing - forecasting floods, landslides and insect pest outbreaks, with a focus on East Africa
[Draft - April 1st, 2015] [Draft 2 - April 6th, 2015] [Draft 2 - April 7, 2015]
Submitted as a theme for a team project at the 2015 International Space Apps Challenge (ISAC) in Tokyo.
ISAC Challenge Category: Earth
Overlap with Earth Challenge themes (日本語翻訳):
The premise of this project is that remote sensing satellites in equatorial low-Earth-orbit (ELEO) could help improve most other LEO satellite remote sensing. The use of remote sensing satellites to help forecast and monitor disasters is of life-and-death importance, especially in tropical mountain regions. These disasters include floods and landslides. Such disasters need better forecasting, monitoring, and relief efforts, and remote sensing satellites are already helping. To improve the lives of the poorest of the poor in the tropics, we should consider not just floods and landslides, but also insect activity. Watershed disasters like floods and mudslides can cause insect pest outbreaks. There is a large base of satellite imagery from around the equator to support research on the links between the sudden watershed disasters and the "slow-motion" disasters of pest outbreaks. The Apps Challenge is a chance to experiment with the needed image processing. The high visitation rates of ELEO satellites might be simulated from the imagery available, to see what can be learned from oversampling. This will mean factoring out the effects of changes in landscape shadowing. (The issues explored need not be limited to those of optical remote sensing, however). One or more software components could be prototyped and demonstrated in the 2015 Apps Challenge weekend. To find the right images for simulated ELEO, we might need a special image-database interface. We will need image processing to demonstrate oversampling. We will need GIS displays of the image processing results. For ground truth, we need mobile apps for environmental data collection (from pest activity to landslide-risk zonation). We might also want mobile apps for field researchers to communicate using mobile directional antennae via the proposed ELEO satellites. The same message system used for day-to-day research could be used by people stranded by a watershed disaster to update relief efforts, when all other communication channels fail.
The premise of this project is that the potential of equatorial low-Earth-orbit (ELEO) remote sensing satellites is still unrealized. Remote sensing satellites that revisit the same points very frequently (every 90 minutes or so) can be smaller. This is because oversampling can mostly compensate for the narrower sensor apertures of smaller satellites. If changes in shadowing (i.e., from the sun’s angle over terrain, and from transitory clouds) can be normalized away, this oversampling should be effective for all but the most rapid environmental changes. Lessons can be learned by correlating the data acquired from frequent overflights by ELEO remote sensing satellites with data from similarly-intensive ground observations, as well as from aerial remote sensing. These lessons might be applied to remote sensing satellites in orbits that cover more latitudes than just the equator, perhaps yielding value equal to the ELEO satellites even with lower visitation rates, and with less intensive supporting observations, both from the ground and from the air.
The use of remote sensing satellites to help forecast and monitor disasters like floods and landslides is literally of vital importance, and especially so in tropical mountain regions. The life-saving potential need not end with basic remote sensing. In the wake of these disasters, such satellites might also serve some emergency communications purposes. For example, if a mobile phone can be positioned at the focal point of an improvised parabolic dish pointed continuously at the satellite as it flies over, the satellite might act as a message buffer when local (terrestrial) links have failed because of damage from storms, floods or landslides. However, such disasters, while tragic, are not very frequent. You'd prefer ELEO remote-sensing satellites to be useful in other ways during normal times. Such satellites will fly just as frequently over an equatorial "ribbon" of tropical ocean, so surely there could also be oceanographic uses for all of the sensors and communications aboard. But the focus here is on the value of an Earth-to-orbit connection for communities at risk in the major land masses.
As part of improving people's lives in densely populated areas in the impoverished tropics, we should consider forecasting, monitoring and supporting disaster relief efforts related not just to floods and landslides, but also as much as possible to insect activity. Individual insects, even swarms, can't be seen at all from space, of course. However, the conditions under which they proliferate, as well as some of the (often disastrous) results of their proliferation, can be monitored to some extent with satellite remote sensing.
There is a relationship here between insect pest outbreaks and the watershed disasters of floods and mudslides. Insect pests can cause "slow-motion disasters" -- crop failures, disease outbreaks, in the same watersheds from which the floods and mudslides come. Droughts that dry up the soil around plant roots and kill soil-anchoring plants can loosen the soil on slopes. This increases landslide risk during a torrential rain, and can mean that downpours carry more mud into flooded areas. But insects can speed this process, by increasing the plant death rate. Droughts can also weaken crop plants and make them more vulnerable to insect pests, contributing to malnutrition, famine and refugee flows. Disastrous flooding in one area can carry dangerous insects (malarial mosquitoes, tsetse flies) to other areas. Climate change makes all of these problems worse. Greater variations in precipitation, to which East African peoples are especially exposed -- combined with high population growth (also a feature of East Africa) can make all of these disasters more lethal.
The Apps Challenge is an opportunity to experiment with the relevant image processing, particularly with simulated oversampling from multiple overflights, and with factoring out the effects of changes in landscape shadowing within a single daylight period. A large base of satellite imagery from equatorial land areas already exists. A great deal of it must overlap geographically, but with different sun angles and different cloud cover. These images can be collected from public sources. There is also a growing base of open source software for GIS and analyzing remote-sensing imagery. There is an established literature on remote sensing as it relates to estimating populations of insect pests.
The Apps Challenge is an opportunity to experiment
(1) with image processing, to simulate what a smaller satellite would "see",
(2) with simulated oversampling from multiple overflights, to better understand how the satellite's vision could be "improved",
(3) with factoring out the effects of changes in landscape shadowing in daylight periods as the sun’s angle changes, and as clouds pass over, to see how much can be learned about determining the conditions of soil and of insect breeding under different climatic conditions.
The issues explored need not be limited to those of optical remote sensing. There is also the "side view": GPS/GNSS radio signal occultation can provide near-real-time readings of precipitation over areas, before and after passing over them. As well, there is the view from outside the visible spectrum: ground-penetrating radar can inform the model with estimates of the moisture content of soils and forest canopies.
The groundwork to be laid in the time between now and the Apps Challenge weekend should consist of identifying public domain data sets, open source software suites, and relevant experts to consult, and arriving at the definition of a project whose difficulty seems just right for completing something useful in a weekend. The desired result might relatively modest.
Among some possible projects (the team will need to choose):
(1) Distributed heterogeneous image database specialized for this theme.
(2) Graphics-intensive demo using mostly-mocked data.
(3) Implementing some of the image processing that would be required.
(4) An East-Africa-friendly mobile app for collecting data about landslides, floods, insect pests.
(5) An East-Africa-friendly mobile app to be used with an improvised parabolic dish for tracking an ELEO satellite and beaming messages to it -- during normal times, sending collected data about watershed terrain and insect pests, and, in the wake of disasters, messages of potential life-saving value for refugees.
-
Rick Fleeter (Brown University/Sapienza University) and Mazlan Othman (formerly Mahathir's space advisor, then UNOOSA) are both veterans of several proposals for equatorial remote sensing satellites.
-
Duccio Rocchini, lead on some open source terrain reconstruction projects - http://gis.cri.fmach.it/rocchini/
-
icipe, a Kenyan entomology research organization concerned chiefly with food security - http://www.icipe.org/ - East Africa is very prone to floods, mudslides, and insect-borne/water-borne diseases (chiefly malaria).
###Software