Taiga under Lidar's Survey
Boreal zone forests constitute one third of all forests on the planet, almost half of the world reserves of wood. About 60% of boreal forests are located in Russia, mostly in Siberia.
Forest "tracking" is necessary, but traditional methods of forest monitoring are very laborious, and extrapolation of space imagery data leads to serious errors in forest inventory. However, today we have a method for direct estimation of forest stand biomass: laser (lidar) sounding with the help of devices installed on the satellites. Reflected laser signal carries information about the height and forest vertical structure.
Thanks to the laser sounding data and ground-based works made by scientists of the V. N. Sukachev Institute of Forest and NASA scientists within a joint project with NASA and in the framework of a program of the RAS Siberian Branch, the map of the above-ground biomass of forest stands of the south of Yenisei meridian has been created. The results of this work showed that lidar sounding can be successfully used to map the above-ground biomass of forest stands and to estimate forest resources in remote and hard-to-reach areas
Boreal zone forests constitute one third of all forests on the planet, including more than two thirds of coniferous forests. In Asia, their southern boundary reaches China and Mongolia; in America, it goes down to Great Lakes; the northernmost “forest island” Ary-Mas is located on the Taimyr Peninsula.
Boreal forests, about 60% of which are located in Russia, have almost half of the world reserves of wood. These figures are impressive but approximate, since they were obtained on the basis of incomplete and old data. Traditional methods of forest monitoring are very laborious, and extrapolation of space imagery data leads to serious errors in forest inventory. However, today we have a method for direct estimation of forest stand biomass: laser (lidar) sounding
Siberian forests are the largest and better part not only of Russian forests but also of the entire boreal belt of the Earth. However, ground-based monitoring of Siberian forest massifs is not easy because many of them are large, remote, and difficult of access.
Forest “tracking” is necessary, because our forests, unlike tropical ones, are a zone of carbon “sink” and accumulation. Here carbon is accumulated in trunks, soil, roots, peats... In this sense, Siberian forests are a real “buffer reservoir” for biosphere, which decrease the concentration rise of carbon dioxide, one of the main greenhouse gases.
YOU CAN’T CHOP WOOD WITHOUT MAKING THE CHIPS FLYToday Russia’s forest resources include forests themselves (776.1 million hectares), territories temporarily not covered with forests (glades, burned-out forest, thinned forest, etc. – 107 million hectares), as well as forest bogs, sands, tundra, and some other objects with a total area of 296 million hectares. The total area of the forest resources constitutes about 70 % of Russia’s territory (Forests of Russia, 2004).
Coniferous species predominate (70 %), with larch and pine being the most widespread ones; about 6 % of the territory is occupied by cedar. The wood reserve of our forests is over 80 billion m3, including more than half of the world reserves of coniferous species. Annual deposition of carbon by Russia’s forest vegetation is estimated as 250 Mt at a total reserve of about 35 Gt (Zamolodchikov, 2004).
Our forests are affected by cutting, fires, insects-pests, diseases, etc. For instance, logging constitutes 130—150 million m3/year, with the proportion of illegal cutting reaching 10—25 %.
On the whole, each year forests die in a territory of 0.2—1.0 million hectares; the major reasons of forest death are fires (50—70 % of the area of dead forests). Even in the protected forest territory (about 2/3 of all forests) from 15,000 to 35,000 forest fires are reported annually. Fire damage (in cubic meters of wood) is estimated as half the cutting volume. In some years, however, it can be much greater.
The forest area affected by pests and diseases is estimated as 1—10 million hectares/year. Fortunately, as in the case of fires, not all affected forests die (Forests of Russia, 2004)
However, already in this century northern landscapes can change from sink zones of carbon into its source: taiga forests are that part of the Earth where maximal warming is observed (and predicted) (IPCC, 2007). Regional climatic models predict further temperature rise in northern zones, which will cause melting of the permafrost and an increase in the discharge of greenhouse gases “conserved” in the frozen layer. In high latitude forests, higher frequency of forest fires is predicted and observed (Kharuk et al., 2008); the probability of appearance of new pest outbreak areas is increasing (for instance, in many territories the population of the Siberian silkmoth Dendrolimus superans sibiricus is limited by the lack of warmth).
It is evidently necessary to trace the phenomena occurring in our forests, in particular, control the pool of carbon and its component – the above-ground biomass of forest stands. But how can this be done?
Forest monitoring involves regular inspections. In practice, however, the greater part of our forests was revised ten and more years ago. Even today, ground-based methods known from old times are used for this purpose. These methods, however, are so laborious that, for instance, in the Asian part of Russia ground-based survey of the forest resources covers no more than half of all forests.
In the 1970s, the so-called photo-statistical method of forest inventory designed to investigate mostly remote forests was developed. The method is based on airborne photography and images obtained by Russian satellites. Space data were interpreted with the use of airborne falsecolor images of key sites whose area did not exceed 5 % of the entire territory under estimation. This allowed using only a small volume of ground-based investigations.
The arsenal of remote sensing used to investigate forests continued to increase, but, unfortunately, not owing to domestic developments. Devices installed on space platforms allowed high-resolution remote sensing of on-ground surfaces in the visible, infrared, and microwave spectral regions. The obtained space images, however, had a serious drawback: with their help one could not determine the vertical structure and height of the forest stand, the key parameters needed for estimation of the above-ground biomass. It has been possible to solve this problem only with laser (lidar) sounding.
Fundamentals of airborne laser sounding of forests were laid in the USSR (Solodukhin et al, 1977). However, aerosurvey is expensive; therefore, modern remote sensing is performed from space platforms.
Laser sounding with the GLAS system allows direct measurement of morphological features of forest stands: the reflected laser signal carries information about the forest vertical structure. Then one can calculate the average diameters of trunks by using allometric relations. On the basis of all these data, the forest reserve and the amount of the above-ground biomass are determined.
The obtained results, however, are point estimates. For interpolation to a large forest territory, a map showing the main classes of forest stands is needed. Unfortunately, the currently available forest maps are obsolete: for instance, the most cited USSR forest map was made in 1991, and the materials on which it is based are even older. Therefore, to update the forest map, the Terra/MODIS system was used: this type of remote sensing is specially designed for investigations of the vegetation cover (Justice et al., 1998).The system called GLAS (Geosciences Laser Altimeter System) onboard the ICESat satellite (NASA) was initially designed for measurements of the Earth’s ice cover height. However, technical characteristics of the system allowed using it to solve “forest” problems as well (Lefsky et al., 2005; Nelson et al., 2009).
GLAS is based on lasers operating in the green (532 nm) and near infrared (1064 nm) spectral regions. Lasers periodically send pulses with a duration of 5 nanoseconds along the satellite flight path. Lidar “footprints” on the Earth’s surface are spots about 70 m in diameter located at a distance of ~170 m.
Enthusiasts say that lying on a grassy plot and patiently looking into the sky allows one to see the green lidar pulse. Incidentally, the “green beam” is used to investigate the atmosphere; near infrared radiation is used to study the Earth’s surface: there is less noise for the sounding beam in this range.
GLAS data – “waveforms” – allow measurement of objects up to 80 m in height; it is evident that all forest stands of the boreal zone fall within this range. The periodicity of lidar survey of a given territory is about three months (Nelson et al., 2009). The major task of researchers is to relate data of waveforms to the parameters of forest stands: height, vertical structure, wood reserve, and biomass
The MODIS scanner installed on the Terra space platform has 36 channels located in the visible and infrared spectral regions. Satellite data are obtained every day (the only limitation is high cloudiness). Ground-based resolution (the pixel size) is from 250 to 1000 m and the survey range is more than 2300 km. This allows using information from Terra/MODIS both for monitoring and survey mapping of forest territories.
Images are interpreted with the use of data from a comparatively small number of “key sites”, where ground-based investigations are made. In principle, for this purpose one could also use data of forest inventory whose periodicity must be, according to the current standards, 10—20 years. However, for economic reasons these time periods are not observed. Therefore, even in our space age researchers have to bridge gaps in forest valuations.
You can’t see everything from above
The purpose of ground-based investigations is to gather information to calibrate space survey data and to interpret “wave forms” in terms of the above-ground biomass. The task of field specialists is to find lidar “footprints” in a forest territory and to determine, at this place, major forest stand parameters (average age, species composition, the height and diameters of trunks, etc.). These measurements must be complemented by a description of the grass-bush tier, the soil cover, and a general characteristic of soils.
Since every “footprint” of the lidar pulse has specific coordinate referencing, with Global Positioning System (GPS) one can find it in the forest stand, enter it, and make the needed measurements. Such ground-based works in the entire Yenisei meridian, from the Sayan ridges to the trans-polar forest-tundra, were made by scientists of the V. N. Sukachev Institute of Forest and NASA scientists within a joint project with NASA and in the framework of a program of the RAS Siberian Branch.
Constructing a map
To map the above-ground biomass of forest stands, it was necessary to establish interrelations between the parameters of waveforms (reflected laser radiation) obtained by the GLAS system and the characteristics of the stands within lidar “footprints”, first of all, with their medium heights.
For this, more than 50 parameters of the lidar signal (the height of peaks, the front steepness, the signal median, etc.) were used; from them, six most informative parameters were chosen. A total of over 66,000 lidar “footprints” were used in the calculations!
As a result, it has become possible to estimate wood reserves with maximal values of about 600 m3/ha. It should be noted that this range covers practically all forest stands of Siberia (with an average volume of about 100 m3/ha). Certainly, formerly one could find stands with a reserve of up to 900 m3/ ha (for instance, in the West Sayan Siberian pine forests); now a reserve of even 150—200 m3/hectare is considered to be sufficient for logging.
However, to create a map it is not enough to obtain local values of biomass within each lidar “footprint.” The next step is interpolation of these “point” estimates to large forest territories. For this, a map showing major forest types and the so-called projective cover (“percentage” of forest land) is needed.
Such a map was created on the basis of survey in the visible and near infrared spectral regions by a MODIS spectroradiometer and a digital model of the relief. The map shows deciduous forests (birch, aspen), evergreen forests (pine, Siberian pine, fir, and spruce), and mixed and larch forests represented by four grades of crown closure.
As a result, along with the above-ground biomass of trees, the forest type and forest land percentage were determined for each lidar “footprint.”
The average reserve values obtained from GLAS/MODIS data are approximately 12 % higher than those obtained by ordinary ground-based monitoring (Shepashenko et al., 1998). However, if we make calculations for the entire mapped forest area, the difference is less than 2 %.
This means that lidar sounding can be successfully used to map the above-ground biomass of forest stands at the subcontinental level and to estimate forest resources in remote and hard-to-reach areas.
FROM THE EXPEDITION DIARY
July 28. The day of “infiltration” from Tura to the larch forests of Zapolyarye. We have a “fifty-fifty” group: three Americans and three Russians. This is not our first joint expedition: our cooperation with the NASA Biospheric Sciences Branch in Greenbelt started in the 90s of the past century.
...A Mi-8 helicopter is heading for the Kochechum upper reaches. There is a cloud mantle above the “point.” Raindrops are oozing through the cabin torch. The copilot throws a smoke bomb to see where to land.
July 30. It rained all night. A peculiarity of northern rivers is abrupt water level variation. The “permafrost” does not absorb rain moisture, and it runs into the river as into a bath. In a day, the river level can rise 1—2 meters. Sometimes we wake up and see that the tent is in water.
July 31. In the north, boats and helicopters are the best transportation means. To save fuel, we have tied all three boats in single file. There are about 500 versts to the return point of Tura.
August 01. The Kochechum is a fast and wild river in its upper reaches. We had to “dive” into the seething waters squeezed by cliff “cheeks.” The rocks are so beautiful that they cry out to be painted; their structure is such that they ask for the geologist’s tool. There is a color variety of minerals on shoals and spits.
August 02. Using the GPS, we find lidar “footprints.” We measure the heights, diameters, and the crown closure of larches. Other species do not survive here. Alder forest and scraggy birches do not count. And the larch growing here is not ordinary – it is the Gmelini larch, whose tolerance for freezing snowstorms is greater than that of the Siberian larch. Larch forests look gorgeous, especially in spring, when crooked branches are covered with a tender-green “haze,” and after the first frosts, when larches turn golden. It is not birch but larch that should be Russia’s symbol!
August 03. We went past an abandoned deer camp. A mounted bridge over the ravine, lopsided carcasses of houses, a dilapidated deer corral; the people and deer left somewhere after the “perestroika.”
August 05. Mosquitoes are annoying. The Americans acknowledge that our repellent is better than foreign sprays. Our “tick suits” are also up to the mark: the entire group is wearing this taiga uniform. Luckily it is not hot; in July there is only one wish - to hide somewhere from the non-setting sun. The level of photosynthetically active radiation (PAR) is the same as in Amazonia.
August 06. We passed the Polar circle. Jon is trying to enter the Internet, where the NASA blog has online reports about our expedition. The computer battery is hopelessly down, and we cannot recharge from solar batteries because of the clouds.
August 07. The Kochechum is cold and swollen from rains. There are shallows here and there… In front of us, billows are rising again, foam and splashes are flying. The river is flowing faster. We are going along the main stream, with two more boats in tow. In front of a shallow, the river falls sharply: the main jet, taking along the near-bank jets, abruptly goes downhill, beats against the rock – and suddenly turns left. In the gorge squeezed by rocks, there are crests of standing waves.
I am trying to catch wave crests of the main stream, muffling the motor at the fall and revving up on the wave rise. I look back quickly: the back boats meet the waves by their noses, taking in water from the crests. When the water chaos has been left behind, Paul, the youngest American, eyes shining with delight, blurts out: the 5B complication category!” And Jon melancholically asks: “Is there much of that ahead?” What shall I answer him? The natives did not say a word about any problems on the Kochechum.
August 09. There are not many wild birds. We see duck flocks running away from the boat; sandpipers of all sizes are cheeping. Waddling northern geese are getting out of the water with dignity.
August 10. Pasha is cleaning a 1943 model carbine. It won’t shoot or, rather, it shoots only once and does not eject shells. Looking at Pasha’s efforts, Jon fatalistically remarks: “Eat me, wild beasts. I’m here, I’m alone, and I’ve no gun.”
From time to time we really come across wolf traces. Luckily, there are no man-eaters among them.
August 12. Burnt-out forest along the banks, both old and fresh: last year the weather was dry. At the burnt-out sites, we fall fire-damaged trees and make saw cuts. Burnt-out trees are a chronicle of fires. However, larch is a pyrophyte, i.e. a species that “likes fires.” Fires drive the frost down, and soils freed from mosses and lichens are again ready to grow larch and new species moving to the north under the influence of climate warming. Among them is cedar, though it grows closer to the south, closer to Tura. Here, cedar will not be able to survive so far.
August 13. Blueberries are scattered along the banks on the sun-lit slopes. These are large, juicy, sourish berries. Jon is humming their popular song: “I had a thrill on the blueberry hill…”
August 16. It is cold, damp, and rainy. We put up tents on the spit among piled boulders. The Americans are moving stones without a murmur to clear the room. Our food supplies are getting short, and all our efforts to get something are fruitless.
August 18. On a dark evening we reach Tura. At the tourist camp, there is a heated bathhouse and a cabin with a cheerfully humming stove and fridge filled with everything we need. And... the non-flying weather into the bargain.
Two days later, the An plane, having forced its way through the hanging cyclone drizzle, brought us to Krasnoyarsk. We arrived just in time: early in the morning our American friends had a flight to Moscow. Goodbye, northern larches. Goodbye, gloomy Kochechum!
Today, lidar sounding is used to estimate not only forests of Siberia but also boreal forests of Europe and North America.
The next step in the development of methods to monitor the land biomass is the “DESDynI” project, in which it is planned to orbit two complementary research instruments: the lidar and the InSAR radar. This will increase the spatial resolution of the lidar to about 25 m, and InSAR will provide “all-weather” “picture” in a survey range of 340 km with a ground-based resolution of about 35 m
An advantage of radar survey is the possibility of direct measurements of the land biomass: forest stands, bushes, and the soil cover. Besides, with its help it will be possible to estimate the hydrological regime of forest territories. Lidar survey will be used to calibrate InSAR data by providing measurements of tree heights and their distribution in the forest canopy.
As a result, implementation of the DESDynI project will greatly decrease ground-based forest estimations. This will allow efficient mapping of wood reserves and tracing of carbon balance not only in boreal forests but also on the entire land surface of the Earth.
Forest resources of Russia. M.: VNIITslesresurs, 2004. 633 p.
Kharuk V. I., Ranson K. J., Fedotova E. V. et al, Radar sounding of taiga forests // Lesovedenie. 2000. no. 5. P. 29—34.
Solodukhin V. I., Zhukov A. Ya., Mazhugin I. N. Possibilities of laser aerial photography of forest profiles // Lesn. Khozyaistvo, 1977. V. 10. P. 53—58